Request strategy: don't launch all 3 requests if not needed

reuse code for listing local blocks
add disk i/o speed limit on integrity check

.drone.yml

@@ -6,103 +6,90 @@ workspace:
   base: /drone/garage
 
 volumes:
-- name: cargo_home
+- name: nix_store
+  host:
+    path: /var/lib/drone/nix
+- name: nix_config
   temp: {}
 
+environment:
+  HOME: /drone/garage
+
 steps:
-  - name: restore-cache
-    image: meltwater/drone-cache:dev
+  - name: setup nix
+    image: nixpkgs/nix:nixos-21.05
     volumes:
-    - name: cargo_home
-      path: /drone/cargo
-    environment:
-      AWS_ACCESS_KEY_ID:
-        from_secret: cache_aws_access_key_id
-      AWS_SECRET_ACCESS_KEY:
-        from_secret: cache_aws_secret_access_key
-    pull: true
-    settings:
-      restore: true
-      archive_format: "gzip"
-      bucket: drone-cache
-      cache_key: '{{ .Repo.Name }}_{{ checksum "Cargo.lock" }}_{{ arch }}_{{ os }}_gzip'
-      region: garage
-      mount:
-        - '/drone/cargo'
-        - 'target'
-      path_style: true
-      endpoint: https://garage.deuxfleurs.fr
-    when:
-      branch:
-      - nonexistent_skip_this_step
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
+    commands:
+      - cp nix/nix.conf /etc/nix/nix.conf
+      - nix-build --no-build-output --no-out-link shell.nix --arg release false  -A inputDerivation
 
   - name: code quality
-    image: superboum/garage_builder_amd64:4
+    image: nixpkgs/nix:nixos-21.05
     volumes:
-    - name: cargo_home
-      path: /drone/cargo
-    environment:
-      CARGO_HOME: /drone/cargo
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
     commands:
-      - cargo fmt -- --check
-      - cargo clippy -- --deny warnings
+      - nix-shell --arg release false --run "cargo fmt -- --check"
+      - nix-shell --arg release false --run "cargo clippy -- --deny warnings"
 
   - name: build
-    image: superboum/garage_builder_amd64:4
+    image: nixpkgs/nix:nixos-21.05
     volumes:
-    - name: cargo_home
-      path: /drone/cargo
-    environment:
-      CARGO_HOME: /drone/cargo
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
     commands:
-      - pwd
-      - cargo build
+      - nix-build --no-build-output --argstr target x86_64-unknown-linux-musl --arg release false --argstr git_version $DRONE_COMMIT
 
-  - name: cargo-test
-    image: superboum/garage_builder_amd64:4
+  - name: unit tests
+    image: nixpkgs/nix:nixos-21.05
     volumes:
-    - name: cargo_home
-      path: /drone/cargo
-    environment:
-      CARGO_HOME: /drone/cargo
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
     commands:
-      - cargo test
-
-  - name: rebuild-cache
-    image: meltwater/drone-cache:dev
-    volumes:
-    - name: cargo_home
-      path: /drone/cargo
-    environment:
-      AWS_ACCESS_KEY_ID:
-        from_secret: cache_aws_access_key_id
-      AWS_SECRET_ACCESS_KEY:
-        from_secret: cache_aws_secret_access_key
-    pull: true
-    settings:
-      rebuild: true
-      archive_format: "gzip"
-      bucket: drone-cache
-      cache_key: '{{ .Repo.Name }}_{{ checksum "Cargo.lock" }}_{{ arch }}_{{ os }}_gzip'
-      region: garage
-      mount:
-        - '/drone/cargo'
-        - 'target'
-      path_style: true
-      endpoint: https://garage.deuxfleurs.fr
-    when:
-      branch:
-      - nonexistent_skip_this_step
+      - |
+        nix-build \
+          --no-build-output \
+          --argstr target x86_64-unknown-linux-musl \
+          --argstr compileMode test
+      - ./result*/bin/garage_api*
+      - ./result*/bin/garage_model*
+      - ./result*/bin/garage_rpc*
+      - ./result*/bin/garage_table*
+      - ./result*/bin/garage_util*
+      - ./result*/bin/garage_web*
+      - ./result*/bin/garage*
 
   - name: smoke-test
-    image: superboum/garage_builder_amd64:4
+    image: nixpkgs/nix:nixos-21.05
     volumes:
-    - name: cargo_home
-      path: /drone/cargo
-    environment:
-      CARGO_HOME: /drone/cargo
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
     commands:
-      - ./script/test-smoke.sh || (cat /tmp/garage.log; false)
+      - nix-build --no-build-output --argstr target x86_64-unknown-linux-musl --arg release false --argstr git_version $DRONE_COMMIT
+      - nix-shell --arg release false --run ./script/test-smoke.sh || (cat /tmp/garage.log; false)
+
+trigger:
+  event:
+  - custom
+  - push
+  - pull_request
+  - tag
+  - cron
+
+node:
+  nix: 1
 
 ---
 kind: pipeline
@@ -137,8 +124,393 @@ steps:
       repo:
         - Deuxfleurs/garage
 
+trigger:
+  event:
+  - custom
+  - push
+  - pull_request
+
+node:
+  nix: 1
+
+---
+kind: pipeline
+type: docker
+name: release-linux-x86_64
+
+volumes:
+- name: nix_store
+  host:
+    path: /var/lib/drone/nix
+- name: nix_config
+  temp: {}
+
+environment:
+  TARGET: x86_64-unknown-linux-musl
+
+steps:
+  - name: setup nix
+    image: nixpkgs/nix:nixos-21.05
+    volumes:
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
+    commands:
+      - cp nix/nix.conf /etc/nix/nix.conf
+      - nix-build --no-build-output --no-out-link shell.nix -A inputDerivation
+
+  - name: build
+    image: nixpkgs/nix:nixos-21.05
+    volumes:
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
+    commands:
+      - nix-build --no-build-output --argstr target $TARGET --arg release true --argstr git_version $DRONE_COMMIT
+
+  - name: integration
+    image: nixpkgs/nix:nixos-21.05
+    volumes:
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
+    commands:
+      - nix-shell --run ./script/test-smoke.sh || (cat /tmp/garage.log; false)
+
+  - name: push static binary
+    image: nixpkgs/nix:nixos-21.05
+    volumes:
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
+    environment:
+      AWS_ACCESS_KEY_ID:
+        from_secret: garagehq_aws_access_key_id
+      AWS_SECRET_ACCESS_KEY:
+        from_secret: garagehq_aws_secret_access_key
+    commands:
+      - nix-shell --arg rust false --arg integration false --run "to_s3"
+
+  - name: docker build and publish
+    image: nixpkgs/nix:nixos-21.05
+    volumes:
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
+    environment:
+      DOCKER_AUTH:
+        from_secret: docker_auth
+      DOCKER_PLATFORM: "linux/amd64"
+      CONTAINER_NAME: "dxflrs/amd64_garage"
+      HOME: "/kaniko"
+    commands:
+      - mkdir -p /kaniko/.docker
+      - echo $DOCKER_AUTH > /kaniko/.docker/config.json
+      - export CONTAINER_TAG=${DRONE_TAG:-$DRONE_COMMIT}
+      - nix-shell --arg rust false --arg integration false --run "to_docker"
+
+
+trigger:
+  event:
+  - promote
+  - cron
+
+node:
+  nix: 1
+
+---
+kind: pipeline
+type: docker
+name: release-linux-i686
+
+volumes:
+- name: nix_store
+  host:
+    path: /var/lib/drone/nix
+- name: nix_config
+  temp: {}
+
+environment:
+  TARGET: i686-unknown-linux-musl
+
+steps:
+  - name: setup nix
+    image: nixpkgs/nix:nixos-21.05
+    volumes:
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
+    commands:
+      - cp nix/nix.conf /etc/nix/nix.conf
+      - nix-build --no-build-output --no-out-link shell.nix -A inputDerivation
+
+  - name: build
+    image: nixpkgs/nix:nixos-21.05
+    volumes:
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
+    commands:
+      - nix-build --no-build-output --argstr target $TARGET --arg release true --argstr git_version $DRONE_COMMIT
+
+  - name: integration
+    image: nixpkgs/nix:nixos-21.05
+    volumes:
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
+    commands:
+      - nix-shell --run ./script/test-smoke.sh || (cat /tmp/garage.log; false)
+
+  - name: push static binary
+    image: nixpkgs/nix:nixos-21.05
+    volumes:
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
+    environment:
+      AWS_ACCESS_KEY_ID:
+        from_secret: garagehq_aws_access_key_id
+      AWS_SECRET_ACCESS_KEY:
+        from_secret: garagehq_aws_secret_access_key
+    commands:
+      - nix-shell --arg rust false --arg integration false --run "to_s3"
+
+  - name: docker build and publish
+    image: nixpkgs/nix:nixos-21.05
+    volumes:
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
+    environment:
+      DOCKER_AUTH:
+        from_secret: docker_auth
+      DOCKER_PLATFORM: "linux/386"
+      CONTAINER_NAME: "dxflrs/386_garage"
+      HOME: "/kaniko"
+    commands:
+      - mkdir -p /kaniko/.docker
+      - echo $DOCKER_AUTH > /kaniko/.docker/config.json
+      - export CONTAINER_TAG=${DRONE_TAG:-$DRONE_COMMIT}
+      - nix-shell --arg rust false --arg integration false --run "to_docker"
+
+trigger:
+  event:
+  - promote
+  - cron
+
+node:
+  nix: 1
+
+---
+kind: pipeline
+type: docker
+name: release-linux-aarch64
+
+volumes:
+- name: nix_store
+  host:
+    path: /var/lib/drone/nix
+- name: nix_config
+  temp: {}
+
+environment:
+  TARGET: aarch64-unknown-linux-musl
+
+steps:
+  - name: setup nix
+    image: nixpkgs/nix:nixos-21.05
+    volumes:
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
+    commands:
+      - cp nix/nix.conf /etc/nix/nix.conf
+      - nix-build --no-build-output --no-out-link ./shell.nix --arg rust false --arg integration false -A inputDerivation
+
+  - name: build
+    image: nixpkgs/nix:nixos-21.05
+    volumes:
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
+    commands:
+      - nix-build --no-build-output --argstr target $TARGET --arg release true --argstr git_version $DRONE_COMMIT
+
+  - name: push static binary
+    image: nixpkgs/nix:nixos-21.05
+    volumes:
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
+    environment:
+      AWS_ACCESS_KEY_ID:
+        from_secret: garagehq_aws_access_key_id
+      AWS_SECRET_ACCESS_KEY:
+        from_secret: garagehq_aws_secret_access_key
+    commands:
+      - nix-shell --arg rust false --arg integration false --run "to_s3"
+
+  - name: docker build and publish
+    image: nixpkgs/nix:nixos-21.05
+    volumes:
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
+    environment:
+      DOCKER_AUTH:
+        from_secret: docker_auth
+      DOCKER_PLATFORM: "linux/arm64"
+      CONTAINER_NAME: "dxflrs/arm64_garage"
+      HOME: "/kaniko"
+    commands:
+      - mkdir -p /kaniko/.docker
+      - echo $DOCKER_AUTH > /kaniko/.docker/config.json
+      - export CONTAINER_TAG=${DRONE_TAG:-$DRONE_COMMIT}
+      - nix-shell --arg rust false --arg integration false --run "to_docker"
+
+trigger:
+  event:
+  - promote
+  - cron
+
+node:
+  nix: 1
+
+---
+kind: pipeline
+type: docker
+name: release-linux-armv6l
+
+volumes:
+- name: nix_store
+  host:
+    path: /var/lib/drone/nix
+- name: nix_config
+  temp: {}
+
+environment:
+  TARGET: armv6l-unknown-linux-musleabihf
+
+steps:
+  - name: setup nix
+    image: nixpkgs/nix:nixos-21.05
+    volumes:
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
+    commands:
+      - cp nix/nix.conf /etc/nix/nix.conf
+      - nix-build --no-build-output --no-out-link --arg rust false --arg integration false -A inputDerivation
+
+  - name: build
+    image: nixpkgs/nix:nixos-21.05
+    volumes:
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
+    commands:
+      - nix-build --no-build-output --argstr target $TARGET --arg release true --argstr git_version $DRONE_COMMIT
+
+  - name: push static binary
+    image: nixpkgs/nix:nixos-21.05
+    volumes:
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
+    environment:
+      AWS_ACCESS_KEY_ID:
+        from_secret: garagehq_aws_access_key_id
+      AWS_SECRET_ACCESS_KEY:
+        from_secret: garagehq_aws_secret_access_key
+    commands:
+      - nix-shell --arg integration false --arg rust false --run "to_s3"
+
+  - name: docker build and publish
+    image: nixpkgs/nix:nixos-21.05
+    volumes:
+    - name: nix_store
+      path: /nix
+    - name: nix_config
+      path: /etc/nix
+    environment:
+      DOCKER_AUTH:
+        from_secret: docker_auth
+      DOCKER_PLATFORM: "linux/arm"
+      CONTAINER_NAME: "dxflrs/arm_garage"
+      HOME: "/kaniko"
+    commands:
+      - mkdir -p /kaniko/.docker
+      - echo $DOCKER_AUTH > /kaniko/.docker/config.json
+      - export CONTAINER_TAG=${DRONE_TAG:-$DRONE_COMMIT}
+      - nix-shell --arg rust false --arg integration false --run "to_docker"
+
+trigger:
+  event:
+  - promote
+  - cron
+
+node:
+  nix: 1
+
+---
+kind: pipeline
+type: docker
+name: refresh-release-page
+
+volumes:
+- name: nix_store
+  host:
+    path: /var/lib/drone/nix
+
+steps:
+  - name: refresh-index
+    image: nixpkgs/nix:nixos-21.05
+    volumes:
+    - name: nix_store
+      path: /nix
+    environment:
+      AWS_ACCESS_KEY_ID:
+        from_secret: garagehq_aws_access_key_id
+      AWS_SECRET_ACCESS_KEY:
+        from_secret: garagehq_aws_secret_access_key
+    commands:
+      - mkdir -p /etc/nix && cp nix/nix.conf /etc/nix/nix.conf
+      - nix-shell --arg integration false --arg rust false --run "refresh_index"
+
+depends_on:
+  - release-linux-x86_64
+  - release-linux-i686
+  - release-linux-aarch64
+  - release-linux-armv6l
+
+trigger:
+  event:
+  - promote
+  - cron
+
+node:
+  nix: 1
+
 ---
 kind: signature
-hmac: e919f8a66d20ebfeeec56b291a8a0fdd59a482601da987fcf533d96d24768744
+hmac: 1c33490cc2902564c4250a409c156683d0d549b8c9d5aee4e46d1bde4e0ccf2c
 
 ...

Dockerfile

@@ -1,10 +1,7 @@
-FROM archlinux:latest
+FROM scratch
 
-RUN mkdir -p /garage/meta
-RUN mkdir -p /garage/data
 ENV RUST_BACKTRACE=1
 ENV RUST_LOG=garage=info
 
-COPY target/release/garage.stripped /garage/garage
-
-CMD /garage/garage server -c /garage/config.toml
+COPY result/bin/garage /
+CMD [ "/garage", "server"]

Makefile

@@ -1,5 +1,13 @@
+.PHONY: doc all release shell
+
 all:
 	clear; cargo build
 
+doc:
+	cd doc/book; mdbook build
+
 release:
-	RUSTFLAGS="-C link-arg=-fuse-ld=lld -C target-cpu=x86-64 -C target-feature=+sse2" cargo build --release --no-default-features
+	nix-build --arg release true
+
+shell:
+	nix-shell

default.nix

@@ -0,0 +1,66 @@
+{
+  system ? builtins.currentSystem,
+  release ? false,
+  target ? "x86_64-unknown-linux-musl",
+  compileMode ? null,
+  git_version ? null,
+}:
+
+with import ./nix/common.nix;
+
+let
+  crossSystem = { config = target; };
+in let
+  pkgs = import pkgsSrc {
+    inherit system crossSystem;
+    overlays = [ cargo2nixOverlay ];
+  };
+
+  /*
+   The following complexity should be abstracted by makePackageSet' (note the final quote).
+   However its code uses deprecated features of rust-overlay that can lead to bug.
+   Instead, we build our own rustChannel object with the recommended API of rust-overlay.
+   */
+  rustChannel = pkgs.rustPlatform.rust;
+
+  overrides = pkgs.buildPackages.rustBuilder.overrides.all ++ [
+    (pkgs.rustBuilder.rustLib.makeOverride {
+      name = "garage";
+      overrideAttrs = drv: if git_version != null then {
+        preConfigure = ''
+          ${drv.preConfigure or ""}
+          export GIT_VERSION="${git_version}"
+        '';
+      } else {};
+    })
+  ];
+
+  packageFun = import ./Cargo.nix;
+
+  rustPkgs = pkgs.rustBuilder.makePackageSet {
+    inherit packageFun rustChannel release;
+    packageOverrides = overrides;
+
+    buildRustPackages = pkgs.buildPackages.rustBuilder.makePackageSet {
+      inherit rustChannel packageFun;
+      packageOverrides = overrides;
+    };
+
+    localPatterns = [
+      /*
+       The way the default rules are written make think we match recursively, on full path, but the rules are misleading.
+       In fact, the regex is only called on root elements of the crate (and not recursively).
+       This behavior does not work well with our nested modules.
+       We tried to build a "deny list" but negative lookup ahead are not supported on Nix.
+       As a workaround, we have to register all our submodules in this allow list...
+       */
+       ''^(src|tests)''          # fixed default
+       ''.*\.(rs|toml)$''        # fixed default
+       ''^(crdt|replication|cli)''   # our crate submodules
+    ];
+  };
+
+in
+  if compileMode == "test"
+    then builtins.mapAttrs (name: value: rustPkgs.workspace.${name} { inherit compileMode; }) rustPkgs.workspace
+    else rustPkgs.workspace.garage { inherit compileMode; }

doc/book/src/SUMMARY.md

@@ -2,23 +2,27 @@
 
 [The Garage Data Store](./intro.md)
 
-- [Getting Started](./getting_started/index.md)
-  - [Get a binary](./getting_started/binary.md)
-  - [Configure the daemon](./getting_started/daemon.md)
-  - [Control the daemon](./getting_started/control.md)
-  - [Configure a cluster](./getting_started/cluster.md)
-  - [Create buckets and keys](./getting_started/bucket.md)
-  - [Handle files](./getting_started/files.md)
+- [Quick start](./quick_start/index.md)
 
 - [Cookbook](./cookbook/index.md)
-  - [Host a website](./cookbook/website.md)
-  - [Integrate as a media backend]()
-  - [Operate a cluster]()
+  - [Building from source](./cookbook/from_source.md)
+  - [Integration with systemd](./cookbook/systemd.md)
+  - [Configuring a reverse proxy](./cookbook/reverse_proxy.md)
+  - [Production Deployment](./cookbook/real_world.md)
   - [Recovering from failures](./cookbook/recovering.md)
 
+- [Integrations](./connect/index.md)
+  - [Apps (Nextcloud, Peertube...)](./connect/apps.md)
+  - [Websites (Hugo, Jekyll, Publii...)](./connect/websites.md)
+  - [Repositories (Docker, Nix...)](./connect/repositories.md)
+  - [CLI tools (rclone, awscli, mc...)](./connect/cli.md)
+  - [Your code (PHP, JS, Go...)](./connect/code.md)
+
+
 - [Reference Manual](./reference_manual/index.md)
-  - [Garage CLI]()
-  - [S3 API](./reference_manual/s3_compatibility.md)
+  - [Garage configuration file](./reference_manual/configuration.md)
+  - [Garage CLI](./reference_manual/cli.md)
+  - [S3 compatibility status](./reference_manual/s3_compatibility.md)
 
 - [Design](./design/index.md)
   - [Related Work](./design/related_work.md)
@@ -26,7 +30,9 @@
 
 - [Development](./development/index.md)
   - [Setup your environment](./development/devenv.md)
-  - [Your first contribution]()
+  - [Development scripts](./development/scripts.md)
+  - [Release process](./development/release_process.md)
 
 - [Working Documents](./working_documents/index.md)
   - [Load Balancing Data](./working_documents/load_balancing.md)
+  - [Migrating from 0.3 to 0.4](./working_documents/migration_04.md)

doc/book/src/connect/apps.md

@@ -0,0 +1,388 @@
+# Apps (Nextcloud, Peertube...)
+
+In this section, we cover the following software: [Nextcloud](#nextcloud), [Peertube](#peertube), [Mastodon](#mastodon), [Matrix](#matrix)
+
+## Nextcloud
+
+Nextcloud is a popular file synchronisation and backup service.
+By default, Nextcloud stores its data on the local filesystem.
+If you want to expand your storage to aggregate multiple servers, Garage is the way to go.
+
+A S3 backend can be configured in two ways on Nextcloud, either as Primary Storage or as an External Storage.
+Primary storage will store all your data on S3, in an opaque manner, and will provide the best performances.
+External storage enable you to select which data will be stored on S3, your file hierarchy will be preserved in S3, but it might be slower.
+
+In the following, we cover both methods but before reading our guide, we suppose you have done some preliminary steps.
+First, we expect you have an already installed and configured Nextcloud instance.
+Second, we suppose you have created a key and a bucket.
+
+As a reminder, you can create a key for your nextcloud instance as follow:
+
+```bash
+garage key new --name nextcloud-key
+```
+
+Keep the Key ID and the Secret key in a pad, they will be needed later.  
+Then you can create a bucket and give read/write rights to your key on this bucket with:
+
+```bash
+garage bucket create nextcloud
+garage bucket allow nextcloud --read --write --key nextcloud-key
+```
+
+
+### Primary Storage
+
+Now edit your Nextcloud configuration file to enable object storage.
+On my installation, the config. file is located at the following path: `/var/www/nextcloud/config/config.php`.  
+We will add a new root key to the `$CONFIG` dictionnary named `objectstore`:
+
+```php
+<?php
+$CONFIG = array(
+/* your existing configuration */
+'objectstore' => [
+    'class' => '\\OC\\Files\\ObjectStore\\S3',
+    'arguments' => [
+        'bucket' => 'nextcloud',   // Your bucket name, must be created before
+        'autocreate' => false,     // Garage does not support autocreate
+        'key'    => 'xxxxxxxxx',   // The Key ID generated previously
+        'secret' => 'xxxxxxxxx',   // The Secret key generated previously
+        'hostname' => '127.0.0.1', // Can also be a domain name, eg. garage.example.com
+        'port' => 3900,            // Put your reverse proxy port or your S3 API port
+        'use_ssl' => false,        // Set it to true if you have a TLS enabled reverse proxy
+        'region' => 'garage',      // Garage has only one region named "garage"
+        'use_path_style' => true   // Garage supports only path style, must be set to true
+    ],
+],
+```
+
+That's all, your Nextcloud will store all your data to S3.
+To test your new configuration, just reload your Nextcloud webpage and start sending data.
+
+*External link:* [Nextcloud Documentation > Primary Storage](https://docs.nextcloud.com/server/latest/admin_manual/configuration_files/primary_storage.html)
+
+### External Storage
+
+**From the GUI.** Activate the "External storage support" app from the "Applications" page (click on your account icon on the top right corner of your screen to display the menu). Go to your parameters page (also located below your account icon). Click on external storage (or the corresponding translation in your language).
+
+[![Screenshot of the External Storage form](./cli-nextcloud-gui.png)](./cli-nextcloud-gui.png)
+*Click on the picture to zoom*
+
+Add a new external storage. Put what you want in "folder name" (eg. "shared"). Select "Amazon S3". Keep "Access Key" for the Authentication field.
+In Configuration, put your bucket name (eg. nextcloud), the host (eg. 127.0.0.1), the port (eg. 3900 or 443), the region (garage). Tick the SSL box if you have put an HTTPS proxy in front of garage. You must tick the "Path access" box and you must leave the "Legacy authentication (v2)" box empty. Put your Key ID (eg. GK...) and your Secret Key in the last two input boxes. Finally click on the tick symbol on the right of your screen.
+
+Now go to your "Files" app and a new "linked folder" has appeared with the name you chose earlier (eg. "shared").
+
+*External link:* [Nextcloud Documentation > External Storage Configuration GUI](https://docs.nextcloud.com/server/latest/admin_manual/configuration_files/external_storage_configuration_gui.html)
+
+**From the CLI.** First install the external storage application:
+
+```bash
+php occ app:install files_external
+```
+
+Then add a new mount point with:
+
+```bash
+ php occ files_external:create \
+  -c bucket=nextcloud \
+  -c hostname=127.0.0.1 \
+  -c port=3900 \
+  -c region=garage \
+  -c use_ssl=false \
+  -c use_path_style=true \
+  -c legacy_auth=false \
+  -c key=GKxxxx \
+  -c secret=xxxx \
+  shared amazons3 amazons3::accesskey
+```
+
+Adapt the `hostname`, `port`, `use_ssl`, `key`, and `secret` entries to your configuration.
+Do not change the `use_path_style` and `legacy_auth` entries, other configurations are not supported.
+
+*External link:* [Nextcloud Documentation > occ command > files external](https://docs.nextcloud.com/server/latest/admin_manual/configuration_server/occ_command.html#files-external-label)
+
+
+## Peertube
+
+Peertube proposes a clever integration of S3 by directly exposing its endpoint instead of proxifying requests through the application.
+In other words, Peertube is only responsible of the "control plane" and offload the "data plane" to Garage.
+In return, this system is a bit harder to configure, especially with Garage that supports less feature than other older S3 backends.
+We show that it is still possible to configure Garage with Peertube, allowing you to spread the load and the bandiwdth usage on the Garage cluster.
+
+### Enable path-style access by patching Peertube
+
+First, you will need to apply a small patch on Peertube ([#4510](https://github.com/Chocobozzz/PeerTube/pull/4510)):
+
+```diff
+From e3b4c641bdf67e07d406a1d49d6aa6b1fbce2ab4 Mon Sep 17 00:00:00 2001
+From: Martin Honermeyer <maze@strahlungsfrei.de>
+Date: Sun, 31 Oct 2021 12:34:04 +0100
+Subject: [PATCH] Allow setting path-style access for object storage
+
+---
+ config/default.yaml                                           | 4 ++++
+ config/production.yaml.example                                | 4 ++++
+ server/initializers/config.ts                                 | 1 +
+ server/lib/object-storage/shared/client.ts                    | 3 ++-
+ .../production/config/custom-environment-variables.yaml       | 2 ++
+ 5 files changed, 13 insertions(+), 1 deletion(-)
+
+diff --git a/config/default.yaml b/config/default.yaml
+index cf9d69a6211..4efd56fb804 100644
+--- a/config/default.yaml
++++ b/config/default.yaml
+@@ -123,6 +123,10 @@ object_storage:
+     # You can also use AWS_SECRET_ACCESS_KEY env variable
+     secret_access_key: ''
+ 
++  # Reference buckets via path rather than subdomain
++  # (i.e. "my-endpoint.com/bucket" instead of "bucket.my-endpoint.com")
++  force_path_style: false
++
+   # Maximum amount to upload in one request to object storage
+   max_upload_part: 2GB
+ 
+diff --git a/config/production.yaml.example b/config/production.yaml.example
+index 70993bf57a3..9ca2de5f4c9 100644
+--- a/config/production.yaml.example
++++ b/config/production.yaml.example
+@@ -121,6 +121,10 @@ object_storage:
+     # You can also use AWS_SECRET_ACCESS_KEY env variable
+     secret_access_key: ''
+ 
++  # Reference buckets via path rather than subdomain
++  # (i.e. "my-endpoint.com/bucket" instead of "bucket.my-endpoint.com")
++  force_path_style: false
++
+   # Maximum amount to upload in one request to object storage
+   max_upload_part: 2GB
+ 
+diff --git a/server/initializers/config.ts b/server/initializers/config.ts
+index 8375bf4304c..d726c59a4b6 100644
+--- a/server/initializers/config.ts
++++ b/server/initializers/config.ts
+@@ -91,6 +91,7 @@ const CONFIG = {
+       ACCESS_KEY_ID: config.get<string>('object_storage.credentials.access_key_id'),
+       SECRET_ACCESS_KEY: config.get<string>('object_storage.credentials.secret_access_key')
+     },
++    FORCE_PATH_STYLE: config.get<boolean>('object_storage.force_path_style'),
+     VIDEOS: {
+       BUCKET_NAME: config.get<string>('object_storage.videos.bucket_name'),
+       PREFIX: config.get<string>('object_storage.videos.prefix'),
+diff --git a/server/lib/object-storage/shared/client.ts b/server/lib/object-storage/shared/client.ts
+index c9a61459336..eadad02f93f 100644
+--- a/server/lib/object-storage/shared/client.ts
++++ b/server/lib/object-storage/shared/client.ts
+@@ -26,7 +26,8 @@ function getClient () {
+         accessKeyId: OBJECT_STORAGE.CREDENTIALS.ACCESS_KEY_ID,
+         secretAccessKey: OBJECT_STORAGE.CREDENTIALS.SECRET_ACCESS_KEY
+       }
+-      : undefined
++      : undefined,
++    forcePathStyle: CONFIG.OBJECT_STORAGE.FORCE_PATH_STYLE
+   })
+ 
+   logger.info('Initialized S3 client %s with region %s.', getEndpoint(), OBJECT_STORAGE.REGION, lTags())
+diff --git a/support/docker/production/config/custom-environment-variables.yaml b/support/docker/production/config/custom-environment-variables.yaml
+index c7cd28e6521..a960bab0bc9 100644
+--- a/support/docker/production/config/custom-environment-variables.yaml
++++ b/support/docker/production/config/custom-environment-variables.yaml
+@@ -54,6 +54,8 @@ object_storage:
+ 
+   region: "PEERTUBE_OBJECT_STORAGE_REGION"
+ 
++  force_path_style: "PEERTUBE_OBJECT_STORAGE_FORCE_PATH_STYLE"
++
+   max_upload_part:
+     __name: "PEERTUBE_OBJECT_STORAGE_MAX_UPLOAD_PART"
+     __format: "json"
+```
+
+You can then recompile it with:
+
+```
+npm run build
+```
+
+And it can be started with:
+
+```
+NODE_ENV=production NODE_CONFIG_DIR=/srv/peertube/config node dist/server.js
+```
+
+
+### Create resources in Garage
+
+Create a key for Peertube:
+
+```bash
+garage key new --name peertube-key
+```
+
+Keep the Key ID and the Secret key in a pad, they will be needed later.  
+
+We need two buckets, one for normal videos (named peertube-video) and one for webtorrent videos (named peertube-playlist).
+```bash
+garage bucket create peertube-video
+garage bucket create peertube-playlist
+```
+
+Now we allow our key to read and write on these buckets:
+
+```
+garage bucket allow peertube-playlist --read --write --key peertube-key
+garage bucket allow peertube-video --read --write --key peertube-key
+```
+
+Finally, we need to expose these buckets publicly to serve their content to users:
+
+```bash
+garage bucket website --allow peertube-playlist
+garage bucket website --allow peertube-video
+```
+
+These buckets are now accessible on the web port (by default 3902) with the following URL: `http://<bucket><root_domain>:<web_port>` where the root domain is defined in your configuration file (by default `.web.garage`). So we have currently the following URLs: 
+  * http://peertube-playlist.web.garage:3902
+  * http://peertube-video.web.garage:3902
+
+Make sure you (will) have a corresponding DNS entry for them.
+
+### Configure a Reverse Proxy to serve CORS
+
+Now we will configure a reverse proxy in front of Garage.
+This is required as we have no other way to serve CORS headers yet.
+For our example, we will use nginx:
+
+```nginx
+server {
+  # In production you should use TLS instead of plain HTTP
+  listen [::]:80;
+
+  server_name peertube-video.web.garage peertube-playlist.web.garage;
+
+  location / {
+    add_header Access-Control-Allow-Origin *;
+    add_header Access-Control-Max-Age 3600;
+    add_header Access-Control-Expose-Headers Content-Length;
+    add_header Access-Control-Allow-Headers Range;
+
+    # We do not forward OPTIONS request to Garage
+    # as it does not know how to interpret them.
+    # Instead, we simply answers 200.
+    if ($request_method !~ ^(GET|HEAD)$ ) {
+      return 200;
+    }
+
+    # If your do not have a Garage instance on the reverse proxy, change the URL here.
+    proxy_pass http://127.0.0.1:3902;
+    proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
+    proxy_set_header Host $host;
+  }
+}
+```
+
+Now make sure that your 2 dns entries are pointing to your reverse proxy.
+
+### Configure Peertube
+
+You must edit the file named `config/production.yaml`, we are only modifying the root key named `object_storage`:
+
+```yaml
+object_storage:
+  enabled: true
+
+  # Put localhost only if you have a garage instance running on that node
+  endpoint: 'http://localhost:3900' # or "garage.example.com" if you have TLS on port 443
+
+  # This entry has been added by our patch, must be set to true
+  force_path_style: true
+
+  # Garage supports only one region for now, named garage
+  region: 'garage'
+
+  credentials:
+    access_key_id: 'GKxxxx'
+    secret_access_key: 'xxxx'
+
+  max_upload_part: 2GB
+
+  streaming_playlists:
+    bucket_name: 'peertube-playlist'
+
+    # Keep it empty for our example
+    prefix: ''
+
+    # You must fill this field to make Peertube use our reverse proxy/website logic
+    base_url: 'http://peertube-playlist.web.garage' # Example: 'https://mirror.example.com'
+
+  # Same settings but for webtorrent videos
+  videos:
+    bucket_name: 'peertube-video'
+    prefix: ''
+    # You must fill this field to make Peertube use our reverse proxy/website logic
+    base_url: 'http://peertube-video.web.garage'
+```
+
+### That's all
+
+Everything must be configured now, simply restart Peertube and try to upload a video.
+You must see in your browser console that data are fetched directly from our bucket (through the reverse proxy).
+
+### Miscellaneous
+
+*Known bug:* The playback does not start and some 400 Bad Request Errors appear in your browser console and on Garage.
+If the description of the error contains HTTP Invalid Range: InvalidRange, the error is due to a buggy ffmpeg version.
+You must avoid the 4.4.0 and use either a newer or older version.
+
+*Associated issues:* [#137](https://git.deuxfleurs.fr/Deuxfleurs/garage/issues/137), [#138](https://git.deuxfleurs.fr/Deuxfleurs/garage/issues/138), [#140](https://git.deuxfleurs.fr/Deuxfleurs/garage/issues/140). These issues are non blocking.
+
+*External link:* [Peertube Documentation > Remote Storage](https://docs.joinpeertube.org/admin-remote-storage)
+
+## Mastodon
+
+https://docs.joinmastodon.org/admin/config/#cdn
+
+## Matrix
+
+### synapse-s3-storage-provider
+
+https://github.com/matrix-org/synapse-s3-storage-provider
+
+### matrix-media-repo
+
+https://github.com/turt2live/matrix-media-repo
+
+## Pixelfed
+
+https://docs.pixelfed.org/technical-documentation/env.html#filesystem
+
+## Pleroma
+
+https://docs-develop.pleroma.social/backend/configuration/cheatsheet/#pleromauploaderss3
+
+## Lemmy
+
+via pict-rs
+https://git.asonix.dog/asonix/pict-rs/commit/f9f4fc63d670f357c93f24147c2ee3e1278e2d97
+
+## Funkwhale
+
+https://docs.funkwhale.audio/admin/configuration.html#s3-storage
+
+## Misskey
+
+https://github.com/misskey-dev/misskey/commit/9d944243a3a59e8880a360cbfe30fd5a3ec8d52d
+
+## Prismo
+
+https://gitlab.com/prismosuite/prismo/-/blob/dev/.env.production.sample#L26-33
+
+## Owncloud Infinite Scale (ocis)
+
+## Unsupported
+
+  - Mobilizon: No S3 integration
+  - WriteFreely: No S3 integration
+  - Plume: No S3 integration

doc/book/src/connect/cli.md

@@ -0,0 +1,127 @@
+# CLI tools
+
+CLI tools allow you to query the S3 API without too many abstractions.
+These tools are particularly suitable for debug, backups, website deployments or any scripted task that need to handle data.
+
+## Minio client (recommended)
+
+Use the following command to set an "alias", i.e. define a new S3 server to be
+used by the Minio client:
+
+```bash
+mc alias set \
+  garage \
+  <endpoint> \
+  <access key> \
+  <secret key> \
+  --api S3v4
+```
+
+Remember that `mc` is sometimes called `mcli` (such as on Arch Linux), to avoid conflicts
+with Midnight Commander.
+
+Some commands:
+
+```bash
+# list buckets
+mc ls garage/
+
+# list objets in a bucket
+mc ls garage/my_files
+
+# copy from your filesystem to garage
+mc cp /proc/cpuinfo garage/my_files/cpuinfo.txt
+
+# copy from garage to your filesystem
+mc cp garage/my_files/cpuinfo.txt /tmp/cpuinfo.txt
+
+# mirror a folder from your filesystem to garage
+mc mirror --overwrite ./book garage/garagehq.deuxfleurs.fr
+```
+
+
+## AWS CLI
+
+Create a file named `~/.aws/credentials` and put:
+
+```toml
+[default]
+aws_access_key_id=xxxx
+aws_secret_access_key=xxxx
+```
+
+Then a file named `~/.aws/config` and put:
+
+```toml
+[default]
+region=garage
+```
+
+Now, supposing Garage is listening on `http://127.0.0.1:3900`, you can list your buckets with:
+
+```bash
+aws --endpoint-url http://127.0.0.1:3900 s3 ls
+```
+
+Passing the `--endpoint-url` parameter to each command is annoying but AWS developers do not provide a corresponding configuration entry.
+As a workaround, you can redefine the aws command by editing the file `~/.bashrc`:
+
+```
+function aws { command aws --endpoint-url http://127.0.0.1:3900 $@ ; }
+```
+
+*Do not forget to run `source ~/.bashrc` or to start a new terminal before running the next commands.*
+
+Now you can simply run:
+
+```bash
+# list buckets
+aws s3 ls
+
+# list objects of a bucket
+aws s3 ls s3://my_files
+
+# copy from your filesystem to garage
+aws s3 cp /proc/cpuinfo s3://my_files/cpuinfo.txt
+
+# copy from garage to your filesystem
+aws s3 cp s3/my_files/cpuinfo.txt /tmp/cpuinfo.txt
+```
+
+## `rclone`
+
+`rclone` can be configured using the interactive assistant invoked using `rclone configure`.
+
+You can also configure `rclone` by writing directly its configuration file.
+Here is a template `rclone.ini` configuration file:
+
+```ini
+[garage]
+type = s3
+provider = Other
+env_auth = false
+access_key_id = <access key>
+secret_access_key = <secret key>
+region = <region>
+endpoint = <endpoint>
+force_path_style = true
+acl = private
+bucket_acl = private
+```
+
+## Cyberduck
+
+TODO
+
+## `s3cmd`
+
+Here is a template for the `s3cmd.cfg` file to talk with Garage:
+
+```ini
+[default]
+access_key = <access key>
+secret_key = <secret key>
+host_base = <endpoint without http(s)://>
+host_bucket = <same as host_base>
+use_https = False | True
+```

doc/book/src/connect/code.md

@@ -0,0 +1 @@
+# Your code (PHP, JS, Go...)

doc/book/src/connect/index.md

@@ -0,0 +1,38 @@
+# Connect it to
+
+To configure an S3 client to interact with Garage, you will need the following
+parameters:
+
+- An **API endpoint**: this corresponds to the HTTP or HTTPS address
+  used to contact the Garage server. When runing Garage locally this will usually
+  be `http://127.0.0.1:3900`. In a real-world setting, you would usually have a reverse-proxy
+  that adds TLS support and makes your Garage server available under a public hostname
+  such as `https://garage.example.com`.
+
+- An **API access key** and its associated **secret key**. These usually look something
+  like this: `GK3515373e4c851ebaad366558` (access key),
+  `7d37d093435a41f2aab8f13c19ba067d9776c90215f56614adad6ece597dbb34` (secret key).
+  These keys are created and managed using the `garage` CLI, as explained in the
+  [quick start](../quick_start/index.md) guide.
+
+Most S3 clients can be configured easily with these parameters,
+provided that you follow the following guidelines:
+
+- **Force path style:** Garage does not support DNS-style buckets, which are now by default
+  on Amazon S3. Instead, Garage uses the legacy path-style bucket addressing.
+  Remember to configure your client to acknowledge this fact.
+
+- **Configuring the S3 region:** Garage requires your client to talk to the correct "S3 region",
+  which is set in the configuration file. This is often set just to `garage`.
+  If this is not configured explicitly, clients usually try to talk to region `us-east-1`.
+  Garage should normally redirect your client to the correct region,
+  but in case your client does not support this you might have to configure it manually.
+
+We will now provide example configurations for the most common clients per category:
+
+  - [Apps](./apps.md)
+  - [Websites](./websites.md)
+  - [Repositories](./repositories.md)
+  - [CLI tools](./cli.md)
+  - [Your code](./code.md)
+

doc/book/src/connect/repositories.md

@@ -0,0 +1 @@
+# Repositories (Docker, Nix...)

doc/book/src/connect/websites.md

@@ -0,0 +1 @@
+# Websites (Hugo, Jekyll, Publii...)

doc/book/src/cookbook/from_source.md

@@ -0,0 +1,51 @@
+# Compiling Garage from source
+
+
+Garage is a standard Rust project.
+First, you need `rust` and `cargo`.
+For instance on Debian:
+
+```bash
+sudo apt-get update
+sudo apt-get install -y rustc cargo
+```
+
+You can also use [Rustup](https://rustup.rs/) to setup a Rust toolchain easily.
+
+## Using source from `crates.io`
+
+Garage's source code is published on `crates.io`, Rust's official package repository.
+This means you can simply ask `cargo` to download and build this source code for you:
+
+```bash
+cargo install garage
+```
+
+That's all, `garage` should be in `$HOME/.cargo/bin`.
+
+You can add this folder to your `$PATH` or copy the binary somewhere else on your system.
+For instance:
+
+```bash
+sudo cp $HOME/.cargo/bin/garage /usr/local/bin/garage
+```
+
+
+## Using source from the Gitea repository
+
+The primary location for Garage's source code is the
+[Gitea repository](https://git.deuxfleurs.fr/Deuxfleurs/garage).
+
+Clone the repository and build Garage with the following commands:
+
+```bash
+git clone https://git.deuxfleurs.fr/Deuxfleurs/garage.git
+cd garage
+cargo build
+```
+
+Be careful, as this will make a debug build of Garage, which will be extremely slow!
+To make a release build, invoke `cargo build --release` (this takes much longer).
+
+The binaries built this way are found in `target/{debug,release}/garage`.
+

doc/book/src/cookbook/index.md

@@ -3,3 +3,23 @@
 A cookbook, when you cook, is a collection of recipes.
 Similarly, Garage's cookbook contains a collection of recipes that are known to works well!
 This chapter could also be referred as "Tutorials" or "Best practices".
+
+- **[Deploying Garage](real_world.md):** This page will walk you through all of the necessary
+  steps to deploy Garage in a real-world setting.
+
+- **[Configuring S3 clients](clients.md):** This page will explain how to configure
+  popular S3 clients to interact with a Garage server.
+
+- **[Hosting a website](website.md):** This page explains how to use Garage
+  to host a static website.
+
+- **[Recovering from failures](recovering.md):** Garage's first selling point is resilience
+  to hardware failures. This section explains how to recover from such a failure in the
+  best possible way.
+
+- **[Building from source](from_source.md):** This page explains how to build Garage from
+  source in case a binary is not provided for your architecture, or if you want to
+  hack with us!
+
+- **[Starting with Systemd](from_source.md):** This page explains how to run Garage
+  as a Systemd service (instead of as a Docker container).

doc/book/src/cookbook/real_world.md

@@ -0,0 +1,273 @@
+# Deploying Garage on a real-world cluster
+
+To run Garage in cluster mode, we recommend having at least 3 nodes.
+This will allow you to setup Garage for three-way replication of your data,
+the safest and most available mode proposed by Garage.
+
+We recommend first following the [quick start guide](../quick_start/index.md) in order
+to get familiar with Garage's command line and usage patterns.
+
+
+
+## Prerequisites
+
+To run a real-world deployment, make sure the following conditions are met:
+
+- You have at least three machines with sufficient storage space available.
+
+- Each machine has a public IP address which is reachable by other machines.
+  Running behind a NAT is likely to be possible but hasn't been tested for the latest version (TODO).
+
+- Ideally, each machine should have a SSD available in addition to the HDD you are dedicating
+  to Garage. This will allow for faster access to metadata and has the potential
+  to drastically reduce Garage's response times.
+
+- This guide will assume you are using Docker containers to deploy Garage on each node. 
+  Garage can also be run independently, for instance as a [Systemd service](systemd.md).
+  You can also use an orchestrator such as Nomad or Kubernetes to automatically manage
+  Docker containers on a fleet of nodes.
+
+Before deploying Garage on your infrastructure, you must inventory your machines.
+For our example, we will suppose the following infrastructure with IPv6 connectivity:
+
+| Location | Name    | IP Address | Disk Space |
+|----------|---------|------------|------------|
+| Paris    | Mercury | fc00:1::1  | 1 To       |
+| Paris    | Venus   | fc00:1::2  | 2 To       |
+| London   | Earth   | fc00:B::1  | 2 To       |
+| Brussels | Mars    | fc00:F::1  | 1.5 To     |
+
+
+
+## Get a Docker image
+
+Our docker image is currently named `lxpz/garage_amd64` and is stored on the [Docker Hub](https://hub.docker.com/r/lxpz/garage_amd64/tags?page=1&ordering=last_updated).
+We encourage you to use a fixed tag (eg. `v0.4.0`) and not the `latest` tag.
+For this example, we will use the latest published version at the time of the writing which is `v0.4.0` but it's up to you
+to check [the most recent versions on the Docker Hub](https://hub.docker.com/r/lxpz/garage_amd64/tags?page=1&ordering=last_updated).
+
+For example:
+
+```
+sudo docker pull lxpz/garage_amd64:v0.4.0
+```
+
+## Deploying and configuring Garage
+
+On each machine, we will have a similar setup,
+especially you must consider the following folders/files:
+
+- `/etc/garage.toml`: Garage daemon's configuration (see below)
+
+- `/var/lib/garage/meta/`: Folder containing Garage's metadata,
+  put this folder on a SSD if possible
+
+- `/var/lib/garage/data/`: Folder containing Garage's data,
+  this folder will be your main data storage and must be on a large storage (e.g. large HDD)
+
+
+A valid `/etc/garage/garage.toml` for our cluster would look as follows:
+
+```toml
+metadata_dir = "/var/lib/garage/meta"
+data_dir = "/var/lib/garage/data"
+
+replication_mode = "3"
+
+rpc_bind_addr = "[::]:3901"
+rpc_public_addr = "<this node's public IP>:3901"
+rpc_secret = "<RPC secret>"
+
+[s3_api]
+s3_region = "garage"
+api_bind_addr = "[::]:3900"
+
+[s3_web]
+bind_addr = "[::]:3902"
+root_domain = ".web.garage"
+index = "index.html"
+```
+
+Check the following for your configuration files:
+
+- Make sure `rpc_public_addr` contains the public IP address of the node you are configuring.
+  This parameter is optional but recommended: if your nodes have trouble communicating with
+  one another, consider adding it.
+
+- Make sure `rpc_secret` is the same value on all nodes. It should be a 32-bytes hex-encoded secret key.
+  You can generate such a key with `openssl rand -hex 32`.
+
+## Starting Garage using Docker
+
+On each machine, you can run the daemon with:
+
+```bash
+docker run \
+  -d \
+  --name garaged \
+  --restart always \
+  --network host \
+  -v /etc/garage.toml:/etc/garage.toml \
+  -v /var/lib/garage/meta:/var/lib/garage/meta \
+  -v /var/lib/garage/data:/var/lib/garage/data \
+  lxpz/garage_amd64:v0.4.0
+```
+
+It should be restarted automatically at each reboot.
+Please note that we use host networking as otherwise Docker containers
+can not communicate with IPv6.
+
+Upgrading between Garage versions should be supported transparently,
+but please check the relase notes before doing so!
+To upgrade, simply stop and remove this container and
+start again the command with a new version of Garage.
+
+## Controling the daemon
+
+The `garage` binary has two purposes:
+  - it acts as a daemon when launched with `garage server`
+  - it acts as a control tool for the daemon when launched with any other command
+
+Ensure an appropriate `garage` binary (the same version as your Docker image) is available in your path.
+If your configuration file is at `/etc/garage.toml`, the `garage` binary should work with no further change.
+
+You can test your `garage` CLI utility by running a simple command such as:
+
+```bash
+garage status
+```
+
+At this point, nodes are not yet talking to one another.
+Your output should therefore look like follows:
+
+```
+Mercury$ garage node-id
+==== HEALTHY NODES ====
+ID                  Hostname  Address           Tag                   Zone  Capacity
+563e1ac825ee3323…   Mercury   [fc00:1::1]:3901  NO ROLE ASSIGNED
+```
+
+
+## Connecting nodes together
+
+When your Garage nodes first start, they will generate a local node identifier
+(based on a public/private key pair).
+
+To obtain the node identifier of a node, once it is generated,
+run `garage node-id`.
+This will print keys as follows:
+
+```bash
+Mercury$ garage node-id
+563e1ac825ee3323aa441e72c26d1030d6d4414aeb3dd25287c531e7fc2bc95d@[fc00:1::1]:3901
+
+Venus$ garage node-id
+86f0f26ae4afbd59aaf9cfb059eefac844951efd5b8caeec0d53f4ed6c85f332@[fc00:1::2]:3901
+
+etc.
+```
+
+You can then instruct nodes to connect to one another as follows:
+
+```bash
+# Instruct Venus to connect to Mercury (this will establish communication both ways)
+Venus$ garage node connect 563e1ac825ee3323aa441e72c26d1030d6d4414aeb3dd25287c531e7fc2bc95d@[fc00:1::1]:3901
+```
+
+You don't nead to instruct all node to connect to all other nodes:
+nodes will discover one another transitively.
+
+Now if your run `garage status` on any node, you should have an output that looks as follows:
+
+```
+==== HEALTHY NODES ====
+ID                  Hostname  Address           Tag                   Zone  Capacity
+563e1ac825ee3323…   Mercury   [fc00:1::1]:3901  NO ROLE ASSIGNED
+86f0f26ae4afbd59…   Venus     [fc00:1::2]:3901  NO ROLE ASSIGNED
+68143d720f20c89d…   Earth     [fc00:B::1]:3901  NO ROLE ASSIGNED
+212f7572f0c89da9…   Mars      [fc00:F::1]:3901  NO ROLE ASSIGNED
+```
+
+## Giving roles to nodes
+
+We will now inform Garage of the disk space available on each node of the cluster
+as well as the zone (e.g. datacenter) in which each machine is located.
+
+For our example, we will suppose we have the following infrastructure
+(Capacity, Identifier and Zone are specific values to Garage described in the following):
+
+| Location | Name    | Disk Space | `Capacity` | `Identifier` | `Zone` |
+|----------|---------|------------|------------|--------------|--------------|
+| Paris    | Mercury | 1 To       | `2`        | `563e`     | `par1`       |
+| Paris    | Venus   | 2 To       | `4`        | `86f0`     | `par1`       |
+| London   | Earth   | 2 To       | `4`        | `6814`     | `lon1`       |
+| Brussels | Mars    | 1.5 To     | `3`        | `212f`     | `bru1`       |
+
+#### Node identifiers
+
+After its first launch, Garage generates a random and unique identifier for each nodes, such as:
+
+```
+563e1ac825ee3323aa441e72c26d1030d6d4414aeb3dd25287c531e7fc2bc95d
+```
+
+Often a shorter form can be used, containing only the beginning of the identifier, like `563e`,
+which identifies the server "Mercury" located in "Paris" according to our previous table.
+
+The most simple way to match an identifier to a node is to run:
+
+```
+garage status
+```
+
+It will display the IP address associated with each node;
+from the IP address you will be able to recognize the node.
+
+#### Zones
+
+Zones are simply a user-chosen identifier that identify a group of server that are grouped together logically.
+It is up to the system administrator deploying Garage to identify what does "grouped together" means.
+
+In most cases, a zone will correspond to a geographical location (i.e. a datacenter).
+Behind the scene, Garage will use zone definition to try to store the same data on different zones,
+in order to provide high availability despite failure of a zone.
+
+#### Capacity
+
+Garage reasons on an abstract metric about disk storage that is named the *capacity* of a node.
+The capacity configured in Garage must be proportional to the disk space dedicated to the node.
+Due to the way the Garage allocation algorithm works, capacity values must
+be **integers**, and must be **as small as possible**, for instance with
+1 representing the size of your smallest server.
+
+Here we chose that 1 unit of capacity = 0.5 To, so that we can express servers of size
+1 To and 2 To, as wel as the intermediate size 1.5 To, with the integer values 2, 4 and
+3 respectively (see table above).
+
+Note that the amount of data stored by Garage on each server may not be strictly proportional to
+its capacity value, as Garage will priorize having 3 copies of data in different zones,
+even if this means that capacities will not be strictly respected. For example in our above examples,
+nodes Earth and Mars will always store a copy of everything each, and the third copy will
+have 66% chance of being stored by Venus and 33% chance of being stored by Mercury.
+
+#### Injecting the topology
+
+Given the information above, we will configure our cluster as follow:
+
+```
+garage node configure -z par1 -c 2 -t mercury 563e
+garage node configure -z par1 -c 4 -t venus 86f0
+garage node configure -z lon1 -c 4 -t earth 6814
+garage node configure -z bru1 -c 3 -t mars 212f
+```
+
+
+## Using your Garage cluster
+
+Creating buckets and managing keys is done using the `garage` CLI,
+and is covered in the [quick start guide](../quick_start/index.md).
+Remember also that the CLI is self-documented thanks to the `--help` flag and
+the `help` subcommand (e.g. `garage help`, `garage key --help`).
+
+Configuring an S3 client to interact with Garage is covered
+[in the next section](clients.md).

doc/book/src/cookbook/recovering.md

@@ -6,14 +6,18 @@ Fear not! For Garage is fully equipped to handle drive failures, in most common
 
 ## A note on availability of Garage
 
-With nodes dispersed in 3 datacenters or more, here are the guarantees Garage provides with the default replication strategy (3 copies of all data, which is the recommended value):
+With nodes dispersed in 3 zones or more, here are the guarantees Garage provides with the 3-way replication strategy (3 copies of all data, which is the recommended replication mode):
 
-- The cluster remains fully functional as long as the machines that fail are in only one datacenter. This includes a whole datacenter going down due to power/Internet outage.
-- No data is lost as long as the machines that fail are in at most two datacenters.
+- The cluster remains fully functional as long as the machines that fail are in only one zone. This includes a whole zone going down due to power/Internet outage.
+- No data is lost as long as the machines that fail are in at most two zones.
 
-Of course this only works if your Garage nodes are correctly configured to be aware of the datacenter in which they are located.
+Of course this only works if your Garage nodes are correctly configured to be aware of the zone in which they are located.
 Make sure this is the case using `garage status` to check on the state of your cluster's configuration.
 
+In case of temporarily disconnected nodes, Garage should automatically re-synchronize
+when the nodes come back up. This guide will deal with recovering from disk failures
+that caused the loss of the data of a node.
+
 
 ## First option: removing a node
 
@@ -92,7 +96,7 @@ Then, replace the broken node by the new one, using:
 
 ```
 garage node configure --replace <old_node_id> \
-		-c <capacity> -d <datacenter> -t <node_tag> <new_node_id>
+		-c <capacity> -z <zone> -t <node_tag> <new_node_id>
 ```
 
 Garage will then start synchronizing all required data on the new node.

doc/book/src/cookbook/reverse_proxy.md

@@ -0,0 +1 @@
+# Configuring a reverse proxy

doc/book/src/cookbook/systemd.md

@@ -0,0 +1,39 @@
+# Starting Garage with systemd instead of Docker
+
+NOTE: This guide is incomplete. Typicall you would also want to create a separate
+Unix user to run Garage.
+
+Make sure you have the Garage binary installed on your system (see [quick start](../quick_start/index.md)), e.g. at `/usr/local/bin/garage`.
+
+Create a file named `/etc/systemd/system/garage.service`:
+
+```toml
+[Unit]
+Description=Garage Data Store
+After=network-online.target
+Wants=network-online.target
+
+[Service]
+Environment='RUST_LOG=garage=info' 'RUST_BACKTRACE=1'
+ExecStart=/usr/local/bin/garage server -c /etc/garage/garage.toml
+
+[Install]
+WantedBy=multi-user.target
+```
+
+To start the service then automatically enable it at boot:
+
+```bash
+sudo systemctl start garage
+sudo systemctl enable garage
+```
+
+To see if the service is running and to browse its logs:
+
+```bash
+sudo systemctl status garage
+sudo journalctl -u garage
+```
+
+If you want to modify the service file, do not forget to run `systemctl daemon-reload`
+to inform `systemd` of your modifications.

doc/book/src/cookbook/website.md

@@ -1 +1,3 @@
-# Host a website
+# Hosting a website
+
+TODO

doc/book/src/development/devenv.md

@@ -1,17 +1,145 @@
 # Setup your development environment
 
-We propose the following quickstart to setup a full dev. environment as quickly as possible:
+Depending on your tastes, you can bootstrap your development environment in a traditional Rust way or through Nix.
 
-  1. Setup a rust/cargo environment. eg. `dnf install rust cargo`
-  2. Install awscli v2 by following the guide [here](https://docs.aws.amazon.com/cli/latest/userguide/install-cliv2.html).
-  3. Run `cargo build` to build the project
-  4. Run `./script/dev-cluster.sh` to launch a test cluster (feel free to read the script)
-  5. Run `./script/dev-configure.sh` to configure your test cluster with default values (same datacenter, 100 tokens)
-  6. Run `./script/dev-bucket.sh` to create a bucket named `eprouvette` and an API key that will be stored in `/tmp/garage.s3`
-  7. Run `source ./script/dev-env-aws.sh` to configure your CLI environment
-  8. You can use `garage` to manage the cluster. Try `garage --help`.
-  9. You can use the `awsgrg` alias to add, remove, and delete files. Try `awsgrg help`, `awsgrg cp /proc/cpuinfo s3://eprouvette/cpuinfo.txt`, or `awsgrg ls s3://eprouvette`. `awsgrg` is a wrapper on the `aws s3` command pre-configured with the previously generated API key (the one in `/tmp/garage.s3`) and localhost as the endpoint.
+## The Nix way
 
-Now you should be ready to start hacking on garage!
+Nix is a generic package manager we use to precisely define our development environment.
+Instructions on how to install it are given on their [Download page](https://nixos.org/download.html).
 
+Check that your installation is working by running the following commands:
 
+```
+nix-shell --version
+nix-build --version
+nix-env   --version
+```
+
+Now, you can clone our git repository (run `nix-env -iA git` if you do not have git yet):
+
+```bash
+git clone https://git.deuxfleurs.fr/Deuxfleurs/garage
+cd garage
+```
+
+*Optionnaly, you can use our nix.conf file to speed up compilations:*
+
+```bash
+sudo mkdir -p /etc/nix
+sudo cp nix/nix.conf /etc/nix/nix.conf
+sudo killall nix-daemon
+```
+
+Now you can enter our nix-shell, all the required packages will be downloaded but they will not pollute your environment outside of the shell:
+
+```bash
+nix-shell
+```
+
+You can use the traditionnal Rust development workflow:
+
+```bash
+cargo build  # compile the project
+cargo run    # execute the project
+cargo test   # run the tests
+cargo fmt    # format the project, run it before any commit!
+cargo clippy # run the linter, run it before any commit!
+```
+
+You can build the project with Nix by running:
+
+```bash
+nix-build
+```
+
+You can parallelize the build (if you use our nix.conf file, it is already automatically done).
+To use all your cores when building a derivation use `-j`, and to build multiple derivations at once use `--max-jobs`.
+The special value `auto` will be replaced by the number of cores of your computer.
+An example:
+
+```bash
+nix-build -j $(nproc) --max-jobs auto
+```
+
+Our build has multiple parameters you might want to set:
+  - `release` build with release optimisations instead of debug
+  - `target allows` for cross compilation
+  - `compileMode` can be set to test or bench to build a unit test runner
+  - `git_version` to inject the hash to display when running `garage stats`
+
+An example:
+
+```bash
+nix-build \
+  --arg release true \
+  --argstr target x86_64-unknown-linux-musl \
+  --argstr compileMode build \
+  --git_version $(git rev-parse HEAD)
+```
+
+*The result is located in `result/bin`. You can pass arguments to cross compile: check `.drone.yml` for examples.*
+
+If you modify a `Cargo.toml` or regenerate any `Cargo.lock`, you must run `cargo2nix`:
+
+```
+cargo2nix -f
+```
+
+Many tools like rclone, `mc` (minio-client), or `aws` (awscliv2) will be available in your environment and will be useful to test Garage.
+
+**This is the recommended method.**
+
+## The Rust way
+
+You need a Rust distribution installed on your computer.
+The most simple way is to install it from [rustup](https://rustup.rs).
+Please avoid using your package manager to install Rust as some tools might be outdated or missing.
+
+Now, check your Rust distribution works by running the following commands:
+
+```bash
+rustc --version
+cargo --version
+rustfmt --version
+clippy-driver --version
+```
+
+Now, you need to clone our git repository ([how to install git](https://git-scm.com/downloads)):
+
+```bash
+git clone https://git.deuxfleurs.fr/Deuxfleurs/garage
+cd garage
+```
+
+You can now use the following commands:
+
+```bash
+cargo build  # compile the project
+cargo run    # execute the project
+cargo test   # run the tests
+cargo fmt    # format the project, run it before any commit!
+cargo clippy # run the linter, run it before any commit!
+```
+
+This is specific to our project, but you will need one last tool, `cargo2nix`.
+To install it, run:
+
+```bash
+cargo install --git https://github.com/superboum/cargo2nix --branch main cargo2nix
+```
+
+You must use it every time you modify a `Cargo.toml` or regenerate a `Cargo.lock` file as follow:
+
+```bash
+cargo build  # Rebuild Cargo.lock if needed
+cargo2nix -f
+```
+
+It will output a `Cargo.nix` file which is a specific `Cargo.lock` file dedicated to Nix that is required by our CI
+which means you must include it in your commits.
+
+Later, to use our scripts and integration tests, you might need additional tools.
+These tools are listed at the end of the `shell.nix` package in the `nativeBuildInputs` part.
+It is up to you to find a way to install the ones you need on your computer.
+
+**A global drawback of this method is that it is up to you to adapt your environment to the one defined in the Nix files.**

doc/book/src/development/miscellaneous_notes.md

@@ -0,0 +1,134 @@
+# Miscellaneous Notes
+
+## Quirks about cargo2nix/rust in Nix
+
+If you use submodules in your crate (like `crdt` and `replication` in `garage_table`), you must list them in `default.nix`
+
+The Windows target does not work. it might be solvable through [overrides](https://github.com/cargo2nix/cargo2nix/blob/master/overlay/overrides.nix). Indeed, we pass `x86_64-pc-windows-gnu` but mingw need `x86_64-w64-mingw32`
+
+We have a simple [PR on cargo2nix](https://github.com/cargo2nix/cargo2nix/pull/201) that fixes critical bugs but the project does not seem very active currently. We must use [my patched version of cargo2nix](https://github.com/superboum/cargo2nix) to enable i686 and armv6l compilation. We might need to contribute to cargo2nix in the future.
+
+
+## Nix
+
+Nix has no armv7 + musl toolchains but armv7l is backward compatible with armv6l.
+
+Signing keys are generated with: 
+
+```
+nix-store --generate-binary-cache-key nix.web.deuxfleurs.fr cache-priv-key.pem cache-pub-key.pem
+```
+
+We copy the secret key in our nix folder:
+
+```
+cp cache-priv-key.pem /etc/nix/signing-key.sec
+```
+
+Manually sign
+    
+We can sign the whole store with:
+
+```
+nix sign-paths --all -k /etc/nix/signing-key.sec
+```
+
+Or simply the current package and its dependencies with:
+
+```
+nix sign-paths --recursive -k /etc/nix/signing-key.sec
+```
+
+Setting a key in `nix.conf` will do the signature at build time automatically without additional commands, edit the `nix.conf` of your builder:
+
+```toml
+secret-key-files = /etc/nix/signing-key.sec
+max-jobs = auto
+cores = 8
+```
+
+Now you are ready to build your packages: 
+
+```bash
+cat > $HOME/.awsrc <<EOF
+export AWS_ACCESS_KEY_ID="xxx"
+export AWS_SECRET_ACCESS_KEY="xxx"
+EOF
+
+# source each time you want to send on the cache
+source ~/.awsrc
+
+# copy garage build dependencies (and not only the output)
+nix-build
+nix-store -qR --include-outputs $(nix-instantiate default.nix) 
+  | xargs nix copy --to 's3://nix?endpoint=garage.deuxfleurs.fr&region=garage'
+  
+# copy shell dependencies
+nix-build shell.nix -A inputDerivation
+nix copy $(nix-store -qR result/) --to 's3://nix?endpoint=garage.deuxfleurs.fr&region=garage' 
+```
+
+More example of nix-copy
+
+```
+# nix-build produces a result/ symlink
+nix copy result/ --to 's3://nix?endpoint=garage.deuxfleurs.fr&region=garage'
+
+# alternative ways to use nix copy
+nix copy nixpkgs.garage --to ...
+nix copy /nix/store/3rbb9qsc2w6xl5xccz5ncfhy33nzv3dp-crate-garage-0.3.0 --to ...
+```
+    
+
+Clear the cache:
+
+```bash
+mc rm --recursive --force garage/nix/
+```
+
+---
+
+A desirable `nix.conf` for a consumer:
+
+```toml
+substituters = https://cache.nixos.org https://nix.web.deuxfleurs.fr
+trusted-public-keys = cache.nixos.org-1:6NCHdD59X431o0gWypbMrAURkbJ16ZPMQFGspcDShjY= nix.web.deuxfleurs.fr:eTGL6kvaQn6cDR/F9lDYUIP9nCVR/kkshYfLDJf1yKs=
+```
+
+And now, whenever you run a command like:
+
+```
+nix-shell
+nix-build
+```
+
+Our cache will be checked.
+
+### Some references about Nix
+
+
+ - https://doc.rust-lang.org/nightly/rustc/platform-support.html
+ - https://nix.dev/tutorials/cross-compilation
+ - https://nixos.org/manual/nix/unstable/package-management/s3-substituter.html
+ - https://fzakaria.com/2020/09/28/nix-copy-closure-your-nix-shell.html
+ - http://www.lpenz.org/articles/nixchannel/index.html
+
+
+## Drone
+
+Do not try to set a build as trusted from the interface or the CLI tool,
+your request would be ignored. Instead, directly edit the database (table `repos`, column `repo_trusted`).
+
+Drone can do parallelism both at the step and the pipeline level. At the step level, parallelism is restricted to the same runner.
+
+## Building Docker containers
+
+We were:
+  - Unable to use the official Docker plugin because
+    - it requires to mount docker socket in the container but it is not recommended
+    - you cant set the platform when building
+  - Unable to use buildah because it needs `CLONE_USERNS` capability
+  - Unable to use the kaniko plugin for Drone as we can't set the target platform
+  - Unable to use the kaniko container provided by Google as we can't run arbitrary logic: we need to put our secret in .docker/config.json.
+
+Finally we chose to build kaniko through nix and use it in a `nix-shell`.

doc/book/src/development/release_process.md

@@ -0,0 +1,195 @@
+# Release process
+
+Before releasing a new version of Garage, our code pass through a succession of checks and transformations.
+We define them as our release process.
+
+## Trigger and classify a release
+
+While we run some tests on every commits, we do not make a release for all of them.
+
+A release can be triggered manually by "promoting" a successful build.
+Otherwise, every weeks, a release build is triggered on the `main` branch.
+
+If the build is from a tag following the regex: `v[0-9]+\.[0-9]+\.[0-9]+`, it will be listed as stable.
+If it is a tag but with a different format, it will be listed as Extra.
+Otherwise, if it is a commit, it will be listed as development.
+This logic is defined in `nix/build_index.nix`.
+
+## Testing
+
+For each commit, we first pass the code to a formatter (rustfmt) and a linter (clippy).
+Then we try to build it in debug mode and run both unit tests and our integration tests.
+
+Additionnaly, when releasing, our integration tests are run on the release build for amd64 and i686.
+
+## Generated Artifacts
+
+We generate the following binary artifacts for now:
+  - **architecture**: amd64, i686, aarch64, armv6
+  - **os**: linux
+  - **format**: static binary, docker container
+
+Additionnaly we also build two web pages:
+  - the documentation (this website)
+  - [the release page](https://garagehq.deuxfleurs.fr/releases.html)
+
+We publish the static binaries on our own garage cluster (you can access them through the releases page)
+and the docker containers on Docker Hub.
+
+## Automation
+
+We automated our release process with Nix and Drone to make it more reliable.
+Here we describe how we have done in case you want to debug or improve it.
+
+### Caching build steps
+
+To speed up the CI, we use the caching feature provided by Nix.
+
+You can benefit from it by using our provided `nix.conf` as recommended or by simply adding the following lines to your file:
+
+```toml
+substituters = https://cache.nixos.org https://nix.web.deuxfleurs.fr
+trusted-public-keys = cache.nixos.org-1:6NCHdD59X431o0gWypbMrAURkbJ16ZPMQFGspcDShjY= nix.web.deuxfleurs.fr:eTGL6kvaQn6cDR/F9lDYUIP9nCVR/kkshYfLDJf1yKs=
+```
+
+Sending to the cache is done through `nix copy`, for example:
+
+```bash
+nix copy --to 's3://nix?endpoint=garage.deuxfleurs.fr&region=garage&secret-key=/etc/nix/signing-key.sec' result
+```
+
+*Note that you need the signing key. In our case, it is stored as a secret in Drone.*
+
+The previous command will only send the built packet and not its dependencies.
+To send its dependency, a tool named `nix-copy-closure` has been created but it is not compatible with the S3 protocol.
+
+Instead, you can use the following commands to list all the runtime dependencies:
+
+```bash
+nix copy \
+  --to 's3://nix?endpoint=garage.deuxfleurs.fr&region=garage&secret-key=/etc/nix/signing-key.sec' \
+  $(nix-store -qR result/)
+```
+
+*We could also write this expression with xargs but this tool is not available in our container.*
+
+But in certain cases, we want to cache compile time dependencies also.
+For example, the Nix project does not provide binaries for cross compiling to i686 and thus we need to compile gcc on our own.
+We do not want to compile gcc each time, so even if it is a compile time dependency, we want to cache it.
+
+This time, the command is a bit more involved:
+
+```bash
+nix copy --to \
+  's3://nix?endpoint=garage.deuxfleurs.fr&region=garage&secret-key=/etc/nix/signing-key.sec' \
+   $(nix-store -qR --include-outputs \
+     $(nix-instantiate))
+```
+
+This is the command we use in our CI as we expect the final binary to change, so we mainly focus on
+caching our development dependencies.
+
+*Currently there is no automatic garbage collection of the cache: we should monitor its growth.
+Hopefully, we can erase it totally without breaking any build, the next build will only be slower.*
+
+In practise, we concluded that we do not want to cache all the compilation dependencies.
+Instead, we want to cache the toolchain we use to build Garage each time we change it.
+So we removed from Drone any automatic update of the cache and instead handle them manually with:
+
+```
+source ~/.awsrc
+nix-shell --run 'refresh_toolchain'
+```
+
+Internally, it will run `nix-build` on  `nix/toolchain.nix` and send the output plus its depedencies to the cache.
+
+To erase the cache:
+
+```
+mc rm --recursive --force 'garage/nix/'
+```
+
+### Publishing Garage
+
+We defined our publishing logic in Nix, mostly as shell hooks.
+You can inspect them in `shell.nix` to see exactly how.
+Here, we will give a quick explanation on how to use them to manually publish a release.
+
+Supposing you just have built garage as follow:
+
+```bash
+nix-build --arg release true
+```
+
+To publish a static binary in `result/bin` on garagehq, run:
+
+```bash
+export AWS_ACCESS_KEY_ID=xxx
+export AWS_SECRET_ACCESS_KEY=xxx
+export DRONE_TAG=handcrafted-1.0.0 # or DRONE_COMMIT
+export TARGET=x86_64-unknown-linux-musl
+
+nix-shell --run to_s3
+```
+
+To create and publish a docker container, run:
+
+```bash
+export DOCKER_AUTH='{ "auths": { "https://index.docker.io/v1/": { "auth": "xxxx" }}}'
+export DOCKER_PLATFORM='linux/amd64' # check GOARCH and GOOS from golang.org
+export CONTAINER_NAME='me/amd64_garage'
+export CONTAINER_TAG='handcrafted-1.0.0'
+
+nix-shell --run to_docker
+```
+
+To rebuild the release page, run:
+```bash
+export AWS_ACCESS_KEY_ID=xxx
+export AWS_SECRET_ACCESS_KEY=xxx
+
+nix-shell --run refresh_index
+```
+
+If you want to compile for different architectures, you will need to repeat all these commands for each architecture.
+
+**In practise, and except for debugging, you will never directly run these commands. Release is handled by drone**
+
+### Drone
+
+Our instance is available at [https://drone.deuxfleurs.fr](https://drone.deuxfleurs.fr).  
+You need an account on [https://git.deuxfleurs.fr](https://git.deuxfleurs.fr) to use it.
+
+**Drone CLI** - Drone has a CLI tool to interact with.
+It can be downloaded from its Github [release page](https://github.com/drone/drone-cli/releases).
+
+To communicate with our instance, you must setup some environment variables.
+You can get them from your [Account Settings](https://drone.deuxfleurs.fr/account).
+
+To make drone easier to use, you could create a `~/.dronerc` that you could source each time you want to use it.
+
+```
+export DRONE_SERVER=https://drone.deuxfleurs.fr
+export DRONE_TOKEN=xxx
+drone info
+```
+
+The CLI tool is very self-discoverable, just append `--help` to each subcommands.
+Start with:
+
+```bash
+drone --help
+```
+
+**.drone.yml** - The builds steps are defined in `.drone.yml`.  
+You can not edit this file without resigning it.
+
+To sign it, you must be a maintainer and then run:
+
+```bash
+drone sign --save Deuxfleurs/garage
+```
+
+Looking at the file, you will see that most of the commands are `nix-shell` and `nix-build` commands with various parameters.
+
+

doc/book/src/development/scripts.md

@@ -0,0 +1,113 @@
+# Development scripts
+
+We maintain a `script/` folder that contains some useful script to ease testing on Garage.
+
+A fully integrated script, `test-smoke.sh`, runs some basic tests on various tools such as minio client, awscli and rclone.
+To run it, enter a `nix-shell` (or install all required tools) and simply run:
+
+```bash
+nix-build # or cargo build
+./script/test-smoke.sh
+```
+
+If something fails, you can find useful logs in `/tmp/garage.log`.
+You can inspect the generated configuration and local data created by inspecting your `/tmp` directory:
+the script creates files and folder prefixed with the name "garage".
+
+## Bootstrapping a test cluster
+
+Under the hood `test-smoke.sh` uses multiple helpers scripts you can also run in case you want to manually test Garage.
+In this section, we introduce 3 scripts to quickly bootstrap a full test cluster with 3 instances.
+
+### 1. Start each daemon
+
+```bash
+./script/dev-cluster.sh
+```
+
+This script spawns 3 Garage instances with 3 configuration files.
+You can inspect the detailed configuration, including ports, by inspecting `/tmp/config.1` (change 1 by the instance number you want).
+
+This script also spawns a simple HTTPS reverse proxy through `socat` for the S3 endpoint that listens on port `4443`.
+Some libraries might require a TLS endpoint to work, refer to our issue [#64](https://git.deuxfleurs.fr/Deuxfleurs/garage/issues/64) for more detailed information on this subject.
+
+This script covers the [Launching the garage server](/quick_start/index.html#launching-the-garage-server) section of our Quick start page.
+
+### 2. Make them join the cluster
+
+```bash
+./script/dev-configure.sh
+```
+
+This script will configure each instance by assigning them a zone (`dc1`) and a weight (`1`).
+
+This script covers the [Configuring your Garage node](/quick_start/index.html#configuring-your-garage-node) section of our Quick start page.
+
+### 3. Create a key and a bucket
+
+```bash
+./script/dev-bucket.sh
+```
+
+This script will create a bucket named `eprouvette` with a key having read and write rights on this bucket.
+The key is stored in a filed named `/tmp/garage.s3` and can be used by the following tools to pre-configure them.
+
+This script covers the [Creating buckets and keys](/quick_start/index.html#creating-buckets-and-keys) section of our Quick start page.
+
+## Handlers for generic tools
+
+We provide wrappers for some CLI tools that configure themselves for your development cluster.
+They are meant to save you some configuration time as to use them, you are only required to source the right file.
+
+### awscli
+
+```bash
+source ./script/dev-env-aws.sh
+
+# some examples
+aws s3 ls s3://eprouvette
+aws s3 cp /proc/cpuinfo s3://eprouvette/cpuinfo.txt
+```
+
+### minio-client
+
+
+```bash
+source ./script/dev-env-mc.sh
+
+# some examples
+mc ls garage/
+mc cp /proc/cpuinfo garage/eprouvette/cpuinfo.txt
+```
+
+### rclone
+
+```bash
+source ./script/dev-env-rclone.sh
+
+# some examples
+rclone lsd garage:
+rclone copy /proc/cpuinfo garage:eprouvette/cpuinfo.txt
+```
+
+### s3cmd
+
+```bash
+source ./script/dev-env-s3cmd.sh
+
+# some examples
+s3cmd ls
+s3cmd put /proc/cpuinfo s3://eprouvette/cpuinfo.txt
+```
+
+### duck
+
+*Warning! Duck is not yet provided by nix-shell.*
+
+```bash
+source ./script/dev-env-duck.sh
+
+# some examples
+duck --list garage:/
+duck --upload garage:/eprouvette/ /proc/cpuinfo
+```

doc/book/src/getting_started/binary.md

@@ -1,44 +0,0 @@
-# Get a binary
-
-Currently, only two installations procedures are supported for Garage: from Docker (x86\_64 for Linux) and from source.
-In the future, we plan to add a third one, by publishing a compiled binary (x86\_64 for Linux).
-We did not test other architecture/operating system but, as long as your architecture/operating system is supported by Rust, you should be able to run Garage (feel free to report your tests!).
-
-## From Docker
-
-Our docker image is currently named `lxpz/garage_amd64` and is stored on the [Docker Hub](https://hub.docker.com/r/lxpz/garage_amd64/tags?page=1&ordering=last_updated).
-We encourage you to use a fixed tag (eg. `v0.2.1`) and not the `latest` tag.
-For this example, we will use the latest published version at the time of the writing which is `v0.2.1` but it's up to you
-to check [the most recent versions on the Docker Hub](https://hub.docker.com/r/lxpz/garage_amd64/tags?page=1&ordering=last_updated).
-
-For example:
-
-```
-sudo docker pull lxpz/garage_amd64:v0.2.1
-```
-
-## From source
-
-Garage is a standard Rust project.
-First, you need `rust` and `cargo`.
-On Debian:
-
-```bash
-sudo apt-get update
-sudo apt-get install -y rustc cargo
-```
-
-Then, you can ask cargo to install the binary for you:
-
-```bash
-cargo install garage
-```
-
-That's all, `garage` should be in `$HOME/.cargo/bin`.
-You can add this folder to your `$PATH` or copy the binary somewhere else on your system.
-For the following, we will assume you copied it in `/usr/local/bin/garage`:
-
-```bash
-sudo cp $HOME/.cargo/bin/garage /usr/local/bin/garage
-```
-

doc/book/src/getting_started/bucket.md

@@ -1,74 +0,0 @@
-# Create buckets and keys
-
-*We use a command named `garagectl` which is in fact an alias you must define as explained in the [Control the daemon](./daemon.md) section.*
-
-In this section, we will suppose that we want to create a bucket named `nextcloud-bucket`
-that will be accessed through a key named `nextcloud-app-key`.
-
-Don't forget that `help` command and `--help` subcommands can help you anywhere, the CLI tool is self-documented! Two examples:
-
-```
-garagectl help
-garagectl bucket allow --help
-```
-
-## Create a bucket
-
-Fine, now let's create a bucket (we imagine that you want to deploy nextcloud):
-
-```
-garagectl bucket create nextcloud-bucket
-```
-
-Check that everything went well:
-
-```
-garagectl bucket list
-garagectl bucket info nextcloud-bucket
-```
-
-## Create an API key
-
-Now we will generate an API key to access this bucket.
-Note that API keys are independent of buckets: one key can access multiple buckets, multiple keys can access one bucket.
-
-Now, let's start by creating a key only for our PHP application:
-
-```
-garagectl key new --name nextcloud-app-key
-```
-
-You will have the following output (this one is fake, `key_id` and `secret_key` were generated with the openssl CLI tool):
-
-```
-Key name: nextcloud-app-key
-Key ID: GK3515373e4c851ebaad366558
-Secret key: 7d37d093435a41f2aab8f13c19ba067d9776c90215f56614adad6ece597dbb34
-Authorized buckets:
-```
-
-Check that everything works as intended:
-
-```
-garagectl key list
-garagectl key info nextcloud-app-key
-```
-
-## Allow a key to access a bucket
-
-Now that we have a bucket and a key, we need to give permissions to the key on the bucket!
-
-```
-garagectl bucket allow \
-  --read \
-  --write 
-  nextcloud-bucket \
-  --key nextcloud-app-key
-```
-
-You can check at any times allowed keys on your bucket with:
-
-```
-garagectl bucket info nextcloud-bucket
-```
-

doc/book/src/getting_started/cluster.md

@@ -1,73 +0,0 @@
-# Configure a cluster
-
-*We use a command named `garagectl` which is in fact an alias you must define as explained in the [Control the daemon](./daemon.md) section.*
-
-In this section, we will inform garage of the disk space available on each node of the cluster
-as well as the site (think datacenter) of each machine.
-
-## Test cluster
-
-As this part is not relevant for a test cluster, you can use this one-liner to create a basic topology:
-
-```bash
-garagectl status | grep UNCONFIGURED | grep -Po '^[0-9a-f]+' | while read id; do 
-    garagectl node configure -d dc1 -c 1 $id
-done
-```
-
-## Real-world cluster
-
-For our example, we will suppose we have the following infrastructure (Capacity, Identifier and Datacenter are specific values to garage described in the following):
-
-| Location | Name    | Disk Space | `Capacity` | `Identifier` | `Datacenter` |
-|----------|---------|------------|------------|--------------|--------------|
-| Paris    | Mercury | 1 To       | `2`        | `8781c5`     | `par1`       |
-| Paris    | Venus   | 2 To       | `4`        | `2a638e`     | `par1`       |
-| London   | Earth   | 2 To       | `4`        | `68143d`     | `lon1`       |
-| Brussels | Mars    | 1.5 To     | `3`        | `212f75`     | `bru1`       |
-
-### Identifier
-
-After its first launch, garage generates a random and unique identifier for each nodes, such as:
-
-```
-8781c50c410a41b363167e9d49cc468b6b9e4449b6577b64f15a249a149bdcbc
-```
-
-Often a shorter form can be used, containing only the beginning of the identifier, like `8781c5`,
-which identifies the server "Mercury" located in "Paris" according to our previous table.
-
-The most simple way to match an identifier to a node is to run:
-
-```
-garagectl status
-```
-
-It will display the IP address associated with each node; from the IP address you will be able to recognize the node.
-
-### Capacity
-
-Garage reasons on an arbitrary metric about disk storage that is named the *capacity* of a node.
-The capacity configured in Garage must be proportional to the disk space dedicated to the node.
-Additionaly, the capacity values used in Garage should be as small as possible, with
-1 ideally representing the size of your smallest server.
-
-Here we chose that 1 unit of capacity = 0.5 To, so that we can express servers of size
-1 To and 2 To, as wel as the intermediate size 1.5 To.
-
-### Datacenter
-
-Datacenter are simply a user-chosen identifier that identify a group of server that are located in the same place.
-It is up to the system administrator deploying garage to identify what does "the same place" means.
-Behind the scene, garage will try to store the same data on different sites to provide high availability despite a data center failure.
-
-### Inject the topology
-
-Given the information above, we will configure our cluster as follow:
-
-```
-garagectl node configure --datacenter par1 -c 2 -t mercury 8781c5
-garagectl node configure --datacenter par1 -c 4 -t venus 2a638e
-garagectl node configure --datacenter lon1 -c 4 -t earth 68143d
-garagectl node configure --datacenter bru1 -c 3 -t mars 212f75
-```

doc/book/src/getting_started/control.md

@@ -1,77 +0,0 @@
-# Control the daemon
-
-The `garage` binary has two purposes:
-  - it acts as a daemon when launched with `garage server ...`
-  - it acts as a control tool for the daemon when launched with any other command
-
-In this section, we will see how to use the `garage` binary as a control tool for the daemon we just started.
-You first need to get a shell having access to this binary, which depends of your configuration:
-  - with `docker-compose`, run `sudo docker-compose exec g1 bash` then `/garage/garage`
-  - with `docker`, run `sudo docker exec -ti garaged bash` then `/garage/garage`
-  - with `systemd`, simply run `/usr/local/bin/garage` if you followed previous instructions
-
-*You can also install the binary on your machine to remotely control the cluster.*
-
-## Talk to the daemon and create an alias
-
-`garage` requires 4 options to talk with the daemon:
-
-```
---ca-cert <ca-cert>            
---client-cert <client-cert>    
---client-key <client-key>      
--h, --rpc-host <rpc-host>
-```
-
-The 3 first ones are certificates and keys needed by TLS, the last one is simply the address of garage's RPC endpoint.
-Because we configure garage directly from the server, we do not need to set `--rpc-host`.
-To avoid typing the 3 first options each time we want to run a command, we will create an alias.
-
-### `docker-compose` alias
-
-```bash
-alias garagectl='/garage/garage \
-  --ca-cert /pki/garage-ca.crt \
-  --client-cert /pki/garage.crt \
-  --client-key /pki/garage.key'
-```
-
-### `docker` alias
-
-```bash
-alias garagectl='/garage/garage \
-  --ca-cert /etc/garage/pki/garage-ca.crt \
-  --client-cert /etc/garage/pki/garage.crt \
-  --client-key /etc/garage/pki/garage.key'
-```
-
-
-### raw binary alias
-
-```bash
-alias garagectl='/usr/local/bin/garage \
-  --ca-cert /etc/garage/pki/garage-ca.crt \
-  --client-cert /etc/garage/pki/garage.crt \
-  --client-key /etc/garage/pki/garage.key'
-```
-
-Of course, if your deployment does not match exactly one of this alias, feel free to adapt it to your needs!
-
-## Test the alias
-
-You can test your alias by running a simple command such as:
-
-```
-garagectl status
-```
-
-You should get something like that as result:
-
-```
-Healthy nodes:
-2a638ed6c775b69a…	37f0ba978d27	[::ffff:172.20.0.101]:3901	UNCONFIGURED/REMOVED
-68143d720f20c89d…	9795a2f7abb5	[::ffff:172.20.0.103]:3901	UNCONFIGURED/REMOVED
-8781c50c410a41b3…	758338dde686	[::ffff:172.20.0.102]:3901	UNCONFIGURED/REMOVED
-```
-
-...which means that you are ready to configure your cluster!

doc/book/src/getting_started/daemon.md

@@ -1,222 +0,0 @@
-# Configure the daemon
-
-Garage is a software that can be run only in a cluster and requires at least 3 instances.
-In our getting started guide, we document two deployment types:
-  - [Test deployment](#test-deployment) though `docker-compose`
-  - [Real-world deployment](#real-world-deployment) through `docker` or `systemd`
-
-In any case, you first need to generate TLS certificates, as traffic is encrypted between Garage's nodes.
-
-## Generating a TLS Certificate
-
-To generate your TLS certificates, run on your machine:
-
-```
-wget https://git.deuxfleurs.fr/Deuxfleurs/garage/raw/branch/master/genkeys.sh
-chmod +x genkeys.sh
-./genkeys.sh
-```
-
-It will creates a folder named `pki` containing the keys that you will used for the cluster.
-
-## Test deployment
-
-Single machine deployment is only described through `docker-compose`.
-
-Before starting, we recommend you create a folder for our deployment:
-
-```bash
-mkdir garage-single
-cd garage-single
-```
-
-We start by creating a file named `docker-compose.yml` describing our network and our containers:
-
-```yml
-version: '3.4'
-
-networks: { virtnet: { ipam: { config: [ subnet: 172.20.0.0/24 ]}}}
-
-services:
-  g1:
-    image: lxpz/garage_amd64:v0.1.1d
-    networks: { virtnet: { ipv4_address: 172.20.0.101 }}
-    volumes:
-      - "./pki:/pki"
-      - "./config.toml:/garage/config.toml"
-
-  g2:
-    image: lxpz/garage_amd64:v0.1.1d
-    networks: { virtnet: { ipv4_address: 172.20.0.102 }}
-    volumes:
-      - "./pki:/pki"
-      - "./config.toml:/garage/config.toml"
-
-  g3:
-    image: lxpz/garage_amd64:v0.1.1d
-    networks: { virtnet: { ipv4_address: 172.20.0.103 }}
-    volumes:
-      - "./pki:/pki"
-      - "./config.toml:/garage/config.toml"
-```
-
-*We define a static network here which is not considered as a best practise on Docker.
-The rational is that Garage only supports IP address and not domain names in its configuration, so we need to know the IP address in advance.*
-
-and then create the `config.toml` file next to it as follow:
-
-```toml
-metadata_dir = "/garage/meta"
-data_dir = "/garage/data"
-rpc_bind_addr = "[::]:3901"
-bootstrap_peers = [
-  "172.20.0.101:3901",
-  "172.20.0.102:3901",
-  "172.20.0.103:3901",
-]
-
-[rpc_tls]
-ca_cert = "/pki/garage-ca.crt"
-node_cert = "/pki/garage.crt"
-node_key = "/pki/garage.key"
-
-[s3_api]
-s3_region = "garage"
-api_bind_addr = "[::]:3900"
-
-[s3_web]
-bind_addr = "[::]:3902"
-root_domain = ".web.garage"
-index = "index.html"
-```
-
-*Please note that we have not mounted `/garage/meta` or `/garage/data` on the host: data will be lost when the container will be destroyed.*
-
-And that's all, you are ready to launch your cluster!
-
-```
-sudo docker-compose up
-```
-
-While your daemons are up, your cluster is still not configured yet.
-However, you can check that your services are still listening as expected by querying them from your host:
-
-```bash
-curl http://172.20.0.{101,102,103}:3902
-```
-
-which should give you:
-
-```
-Not found
-Not found
-Not found
-```
-
-That's all, you are ready to [configure your cluster!](./cluster.md).
-
-## Real-world deployment
-
-Before deploying garage on your infrastructure, you must inventory your machines.
-For our example, we will suppose the following infrastructure:
-
-| Location | Name    | IP Address | Disk Space |
-|----------|---------|------------|------------|
-| Paris    | Mercury | fc00:1::1  | 1 To       |
-| Paris    | Venus   | fc00:1::2  | 2 To       |
-| London   | Earth   | fc00:B::1  | 2 To       |
-| Brussels | Mars    | fc00:F::1  | 1.5 To     |
-
-On each machine, we will have a similar setup, especially you must consider the following folders/files:
-  - `/etc/garage/pki`: Garage certificates, must be generated on your computer and copied on the servers
-  - `/etc/garage/config.toml`: Garage daemon's configuration (defined below)
-  - `/etc/systemd/system/garage.service`: Service file to start garage at boot automatically (defined below, not required if you use docker)
-  - `/var/lib/garage/meta`: Contains Garage's metadata, put this folder on a SSD if possible
-  - `/var/lib/garage/data`: Contains Garage's data, this folder will grows and must be on a large storage, possibly big HDDs.
-
-A valid `/etc/garage/config.toml` for our cluster would be:
-
-```toml
-metadata_dir = "/var/lib/garage/meta"
-data_dir = "/var/lib/garage/data"
-rpc_bind_addr = "[::]:3901"
-bootstrap_peers = [
-  "[fc00:1::1]:3901",
-  "[fc00:1::2]:3901",
-  "[fc00:B::1]:3901",
-  "[fc00:F::1]:3901",
-]
-
-[rpc_tls]
-ca_cert = "/etc/garage/pki/garage-ca.crt"
-node_cert = "/etc/garage/pki/garage.crt"
-node_key = "/etc/garage/pki/garage.key"
-
-[s3_api]
-s3_region = "garage"
-api_bind_addr = "[::]:3900"
-
-[s3_web]
-bind_addr = "[::]:3902"
-root_domain = ".web.garage"
-index = "index.html"
-```
-
-Please make sure to change `bootstrap_peers` to **your** IP addresses!
-
-### For docker users
-
-On each machine, you can run the daemon with:
-
-```bash
-docker run \
-  -d \
-  --name garaged \
-  --restart always \
-  --network host \
-  -v /etc/garage/pki:/etc/garage/pki \
-  -v /etc/garage/config.toml:/garage/config.toml \
-  -v /var/lib/garage/meta:/var/lib/garage/meta \
-  -v /var/lib/garage/data:/var/lib/garage/data \
-  lxpz/garage_amd64:v0.1.1d
-```
-
-It should be restart automatically at each reboot.
-Please note that we use host networking as otherwise Docker containers can no communicate with IPv6.
-
-To upgrade, simply stop and remove this container and start again the command with a new version of garage.
-
-### For systemd/raw binary users
-
-Create a file named `/etc/systemd/system/garage.service`:
-
-```toml
-[Unit]
-Description=Garage Data Store
-After=network-online.target
-Wants=network-online.target
-
-[Service]
-Environment='RUST_LOG=garage=info' 'RUST_BACKTRACE=1'
-ExecStart=/usr/local/bin/garage server -c /etc/garage/config.toml
-
-[Install]
-WantedBy=multi-user.target
-```
-
-To start the service then automatically enable it at boot:
-
-```bash
-sudo systemctl start garage
-sudo systemctl enable garage
-```
-
-To see if the service is running and to browse its logs:
-
-```bash
-sudo systemctl status garage
-sudo journalctl -u garage
-```
-
-If you want to modify the service file, do not forget to run `systemctl daemon-reload`
-to inform `systemd` of your modifications.

doc/book/src/getting_started/files.md

@@ -1,42 +0,0 @@
-# Handle files
-
-We recommend the use of MinIO Client to interact with Garage files (`mc`).
-Instructions to install it and use it are provided on the [MinIO website](https://docs.min.io/docs/minio-client-quickstart-guide.html).
-Before reading the following, you need a working `mc` command on your path.
-
-## Configure `mc`
-
-You need your access key and secret key created in the [previous section](bucket.md).
-You also need to set the endpoint: it must match the IP address of one of the node of the cluster and the API port (3900 by default).
-For this whole configuration, you must set an alias name: we chose `my-garage`, that you will used for all commands.
-
-Adapt the following command accordingly and run it:
-
-```bash
-mc alias set \
-  my-garage \
-  http://172.20.0.101:3900 \
-  <access key> \
-  <secret key> \
-  --api S3v4
-```
-
-You must also add an environment variable to your configuration to inform MinIO of our region (`garage` by default).
-The best way is to add the following snippet to your `$HOME/.bash_profile` or `$HOME/.bashrc` file:
-
-```bash
-export MC_REGION=garage
-```
-
-## Use `mc`
-
-You can not list buckets from `mc` currently.
-
-But the following commands and many more should work:
-
-```bash
-mc cp image.png my-garage/nextcloud-bucket
-mc cp my-garage/nextcloud-bucket/image.png .
-mc ls my-garage/nextcloud-bucket
-mc mirror localdir/ my-garage/another-bucket
-```

doc/book/src/getting_started/index.md

@@ -1,5 +0,0 @@
-# Getting Started
-
-Let's start your Garage journey!
-In this chapter, we explain how to deploy a simple garage cluster and start interacting with it.
-Our goal is to introduce you to Garage's workflows.

doc/book/src/intro.md

@@ -5,7 +5,8 @@
 </p>
 
 <p align="center" style="text-align:center;">
-	[ <a href="https://git.deuxfleurs.fr/Deuxfleurs/garage">Git repository</a>
+	[ <a href="https://garagehq.deuxfleurs.fr/_releases.html">Download</a>
+	| <a href="https://git.deuxfleurs.fr/Deuxfleurs/garage">Git repository</a>
 	| <a href="https://matrix.to/#/%23garage:deuxfleurs.fr">Matrix channel</a>
 	| <a href="https://drone.deuxfleurs.fr/Deuxfleurs/garage">Drone CI</a>
 	]
@@ -70,7 +71,7 @@ We also do not classify Swift as *Simple*.
 **[Ceph](https://ceph.io/ceph-storage/object-storage/):**
 This review holds for the whole Ceph stack, including the RADOS paper, Ceph Object Storage module, the RADOS Gateway, etc.
 At its core, Ceph has been designed to provide *POSIX/Filesystem compatibility* which requires strong consistency, which in turn
-makes Ceph latency-sensitive and fails our *Internet enabled* goal. 
+makes Ceph latency-sensitive and fails our *Internet enabled* goal.
 Due to its industry oriented design, Ceph is also far from being *Simple* to operate and from being *Self-contained & lightweight* which makes it hard to integrate it in an hyperconverged infrastructure.
 In a certain way, Ceph and MinIO are closer together than they are from Garage or OpenStack Swift.
 
@@ -111,3 +112,18 @@ Our code repository and issue tracker, which is the place where you should repor
 
 Garage's source code, is released under the [AGPL v3 License](https://www.gnu.org/licenses/agpl-3.0.en.html).
 Please note that if you patch Garage and then use it to provide any service over a network, you must share your code!
+
+# Funding
+
+The Deuxfleurs association has received a grant from [NGI POINTER](https://pointer.ngi.eu/), to fund 3 people working on Garage full-time for a year: from October 2021 to September 2022.
+
+<div style="display: flex; justify-content: space-around">
+  <a href="https://pointer.ngi.eu/">
+    <img style="height:100px" src="img/ngi-logo.png" alt="NGI Pointer logo">
+  </a>
+  <a href="https://ec.europa.eu/programmes/horizon2020/what-horizon-2020">
+    <img style="height:100px" src="img/eu-flag-logo.png" alt="EU flag logo">
+  </a>
+</div>
+
+_This project has received funding from the European Union’s Horizon 2020 research and innovation programme within the framework of the NGI-POINTER Project funded under grant agreement N° 871528._

doc/book/src/quick_start/index.md

@@ -0,0 +1,273 @@
+# Quick Start
+
+Let's start your Garage journey!
+In this chapter, we explain how to deploy Garage as a single-node server
+and how to interact with it.
+
+Our goal is to introduce you to Garage's workflows.
+Following this guide is recommended before moving on to
+[configuring a real-world deployment](../cookbook/real_world.md).
+
+Note that this kind of deployment should not be used in production, as it provides
+no redundancy for your data!
+
+## Get a binary
+
+Download the latest Garage binary from the release pages on our repository:
+
+<https://git.deuxfleurs.fr/Deuxfleurs/garage/releases>
+
+Place this binary somewhere in your `$PATH` so that you can invoke the `garage`
+command directly (for instance you can copy the binary in `/usr/local/bin`
+or in `~/.local/bin`).
+
+If a binary of the last version is not available for your architecture,
+you can [build Garage from source](../cookbook/from_source.md).
+
+
+## Writing a first configuration file
+
+This first configuration file should allow you to get started easily with the simplest
+possible Garage deployment.
+**Save it as `/etc/garage.toml`.**
+You can also store it somewhere else, but you will have to specify `-c path/to/garage.toml`
+at each invocation of the `garage` binary (for example: `garage -c ./garage.toml server`, `garage -c ./garage.toml status`).
+
+```toml
+metadata_dir = "/tmp/meta"
+data_dir = "/tmp/data"
+
+replication_mode = "none"
+
+rpc_bind_addr = "[::]:3901"
+rpc_public_addr = "127.0.0.1:3901"
+rpc_secret = "1799bccfd7411eddcf9ebd316bc1f5287ad12a68094e1c6ac6abde7e6feae1ec"
+
+bootstrap_peers = []
+
+[s3_api]
+s3_region = "garage"
+api_bind_addr = "[::]:3900"
+
+[s3_web]
+bind_addr = "[::]:3902"
+root_domain = ".web.garage"
+index = "index.html"
+```
+
+The `rpc_secret` value provided above is just an example. It will work, but in
+order to secure your cluster you will need to use another one. You can generate
+such a value with `openssl rand -hex 32`.
+
+
+As you can see in the `metadata_dir` and `data_dir` parameters, we are saving Garage's data
+in `/tmp` which gets erased when your system reboots. This means that data stored on this
+Garage server will not be persistent. Change these to locations on your local disk if you want
+your data to be persisted properly.
+
+
+## Launching the Garage server
+
+Use the following command to launch the Garage server with our configuration file:
+
+```
+RUST_LOG=garage=info garage server
+```
+
+You can tune Garage's verbosity as follows (from less verbose to more verbose):
+
+```
+RUST_LOG=garage=info garage server
+RUST_LOG=garage=debug garage server
+RUST_LOG=garage=trace garage server
+```
+
+Log level `info` is recommended for most use cases.
+Log level `debug` can help you check why your S3 API calls are not working.
+
+
+## Checking that Garage runs correctly
+
+The `garage` utility is also used as a CLI tool to configure your Garage deployment.
+It uses values from the TOML configuration file to find the Garage daemon running on the
+local node, therefore if your configuration file is not at `/etc/garage.toml` you will
+again have to specify `-c path/to/garage.toml`.
+
+If the `garage` CLI is able to correctly detect the parameters of your local Garage node,
+the following command should be enough to show the status of your cluster:
+
+```
+garage status
+```
+
+This should show something like this:
+
+```
+==== HEALTHY NODES ====
+ID                 Hostname  Address         Tag                   Zone  Capacity
+563e1ac825ee3323…  linuxbox  127.0.0.1:3901  NO ROLE ASSIGNED
+```
+
+## Configuring your Garage node
+
+Configuring the nodes in a Garage deployment means informing Garage
+of the disk space available on each node of the cluster
+as well as the zone (e.g. datacenter) each machine is located in.
+
+For our test deployment, we are using only one node. The way in which we configure
+it does not matter, you can simply write:
+
+```bash
+garage node configure -z dc1 -c 1 <node_id>
+```
+
+where `<node_id>` corresponds to the identifier of the node shown by `garage status` (first column).
+You can enter simply a prefix of that identifier.
+For instance here you could write just `garage node configure -z dc1 -c 1 563e`.
+
+
+
+
+## Creating buckets and keys
+
+In this section, we will suppose that we want to create a bucket named `nextcloud-bucket`
+that will be accessed through a key named `nextcloud-app-key`.
+
+Don't forget that `help` command and `--help` subcommands can help you anywhere,
+the CLI tool is self-documented! Two examples:
+
+```
+garage help
+garage bucket allow --help
+```
+
+#### Create a bucket
+
+Let's take an example where we want to deploy NextCloud using Garage as the
+main data storage.
+
+First, create a bucket with the following command:
+
+```
+garage bucket create nextcloud-bucket
+```
+
+Check that everything went well:
+
+```
+garage bucket list
+garage bucket info nextcloud-bucket
+```
+
+#### Create an API key
+
+The `nextcloud-bucket` bucket now exists on the Garage server,
+however it cannot be accessed until we add an API key with the proper access rights.
+
+Note that API keys are independent of buckets:
+one key can access multiple buckets, multiple keys can access one bucket.
+
+Create an API key using the following command:
+
+```
+garage key new --name nextcloud-app-key
+```
+
+The output should look as follows:
+
+```
+Key name: nextcloud-app-key
+Key ID: GK3515373e4c851ebaad366558
+Secret key: 7d37d093435a41f2aab8f13c19ba067d9776c90215f56614adad6ece597dbb34
+Authorized buckets:
+```
+
+Check that everything works as intended:
+
+```
+garage key list
+garage key info nextcloud-app-key
+```
+
+#### Allow a key to access a bucket
+
+Now that we have a bucket and a key, we need to give permissions to the key on the bucket:
+
+```
+garage bucket allow \
+  --read \
+  --write 
+  nextcloud-bucket \
+  --key nextcloud-app-key
+```
+
+You can check at any time the allowed keys on your bucket with:
+
+```
+garage bucket info nextcloud-bucket
+```
+
+
+## Uploading and downlading from Garage
+
+We recommend the use of MinIO Client to interact with Garage files (`mc`).
+Instructions to install it and use it are provided on the
+[MinIO website](https://docs.min.io/docs/minio-client-quickstart-guide.html).
+Before reading the following, you need a working `mc` command on your path.
+
+Note that on certain Linux distributions such as Arch Linux, the Minio client binary
+is called `mcli` instead of `mc` (to avoid name clashes with the Midnight Commander).
+
+#### Configure `mc`
+
+You need your access key and secret key created above.
+We will assume you are invoking `mc` on the same machine as the Garage server,
+your S3 API endpoint is therefore `http://127.0.0.1:3900`.
+For this whole configuration, you must set an alias name: we chose `my-garage`, that you will used for all commands.
+
+Adapt the following command accordingly and run it:
+
+```bash
+mc alias set \
+  my-garage \
+  http://127.0.0.1:3900 \
+  <access key> \
+  <secret key> \
+  --api S3v4
+```
+
+You must also add an environment variable to your configuration to
+inform MinIO of our region (`garage` by default, corresponding to the `s3_region` parameter
+in the configuration file).
+The best way is to add the following snippet to your `$HOME/.bash_profile`
+or `$HOME/.bashrc` file:
+
+```bash
+export MC_REGION=garage
+```
+
+#### Use `mc`
+
+You can not list buckets from `mc` currently.
+
+But the following commands and many more should work:
+
+```bash
+mc cp image.png my-garage/nextcloud-bucket
+mc cp my-garage/nextcloud-bucket/image.png .
+mc ls my-garage/nextcloud-bucket
+mc mirror localdir/ my-garage/another-bucket
+```
+
+
+#### Other tools for interacting with Garage
+
+The following tools can also be used to send and recieve files from/to Garage:
+
+- the [AWS CLI](https://aws.amazon.com/cli/)
+- [`rclone`](https://rclone.org/)
+- [Cyberduck](https://cyberduck.io/)
+- [`s3cmd`](https://s3tools.org/s3cmd)
+
+Refer to the ["configuring clients"](../cookbook/clients.md) page to learn how to configure
+these clients to interact with a Garage server.

doc/book/src/reference_manual/cli.md

@@ -0,0 +1,4 @@
+# Garage CLI
+
+The Garage CLI is mostly self-documented. Make use of the `help` subcommand
+and the `--help` flag to discover all available options.

doc/book/src/reference_manual/configuration.md

@@ -0,0 +1,201 @@
+# Garage configuration file format reference
+
+Here is an example `garage.toml` configuration file that illustrates all of the possible options:
+
+```toml
+metadata_dir = "/var/lib/garage/meta"
+data_dir = "/var/lib/garage/data"
+
+block_size = 1048576
+
+replication_mode = "3"
+
+rpc_secret = "4425f5c26c5e11581d3223904324dcb5b5d5dfb14e5e7f35e38c595424f5f1e6"
+rpc_bind_addr = "[::]:3901"
+rpc_public_addr = "[fc00:1::1]:3901"
+
+bootstrap_peers = [
+    "563e1ac825ee3323aa441e72c26d1030d6d4414aeb3dd25287c531e7fc2bc95d@[fc00:1::1]:3901",
+    "86f0f26ae4afbd59aaf9cfb059eefac844951efd5b8caeec0d53f4ed6c85f332[fc00:1::2]:3901",
+    "681456ab91350f92242e80a531a3ec9392cb7c974f72640112f90a600d7921a4@[fc00:B::1]:3901",
+    "212fd62eeaca72c122b45a7f4fa0f55e012aa5e24ac384a72a3016413fa724ff@[fc00:F::1]:3901",
+]
+
+consul_host = "consul.service"
+consul_service_name = "garage-daemon"
+
+sled_cache_capacity = 134217728
+sled_flush_every_ms = 2000
+
+[s3_api]
+api_bind_addr = "[::]:3900"
+s3_region = "garage"
+
+[s3_web]
+bind_addr = "[::]:3902"
+root_domain = ".web.garage"
+index = "index.html"
+```
+
+The following gives details about each available configuration option.
+
+## Available configuration options
+
+#### `metadata_dir`
+
+The directory in which Garage will store its metadata. This contains the node identifier,
+the network configuration and the peer list, the list of buckets and keys as well
+as the index of all objects, object version and object blocks.
+
+Store this folder on a fast SSD drive if possible to maximize Garage's performance.
+
+#### `data_dir`
+
+The directory in which Garage will store the data blocks of objects.
+This folder can be placed on an HDD. The space available for `data_dir`
+should be counted to determine a node's capacity
+when [configuring it](../getting_started/05_cluster.md).
+
+#### `block_size`
+
+Garage splits stored objects in consecutive chunks of size `block_size`
+(except the last one which might be smaller). The default size is 1MB and
+should work in most cases.  If you are interested in tuning this, feel free
+to do so (and remember to report your findings to us!). If this value is
+changed for a running Garage installation, only files newly uploaded will be
+affected. Previously uploaded files will remain available. This however
+means that chunks from existing files will not be deduplicated with chunks
+from newly uploaded files, meaning you might use more storage space that is
+optimally possible.
+
+#### `replication_mode`
+
+Garage supports the following replication modes:
+
+- `none` or `1`: data stored on Garage is stored on a single node. There is no redundancy,
+  and data will be unavailable as soon as one node fails or its network is disconnected.
+  Do not use this for anything else than test deployments.
+
+- `2`: data stored on Garage will be stored on two different nodes, if possible in different
+  zones. Garage tolerates one node failure before losing data. Data should be available
+  read-only when one node is down, but write operations will fail.
+  Use this only if you really have to.
+
+- `3`: data stored on Garage will be stored on three different nodes, if possible each in
+  a different zones.
+  Garage tolerates two node failure before losing data. Data should be available
+  read-only when two nodes are down, and writes should be possible if only a single node
+  is down.
+
+Note that in modes `2` and `3`,
+if at least the same number of zones are available, an arbitrary number of failures in 
+any given zone is tolerated as copies of data will be spread over several zones.
+
+**Make sure `replication_mode` is the same in the configuration files of all nodes.
+Never run a Garage cluster where that is not the case.**
+
+Changing the `replication_mode` of a cluster might work (make sure to shut down all nodes
+and changing it everywhere at the time), but is not officially supported.
+
+#### `rpc_secret`
+
+Garage uses a secret key that is shared between all nodes of the cluster
+in order to identify these nodes and allow them to communicate together.
+This key should be specified here in the form of a 32-byte hex-encoded
+random string. Such a string can be generated with a command
+such as `openssl rand -hex 32`.
+
+#### `rpc_bind_addr`
+
+The address and port on which to bind for inter-cluster communcations
+(reffered to as RPC for remote procedure calls).
+The port specified here should be the same one that other nodes will used to contact
+the node, even in the case of a NAT: the NAT should be configured to forward the external
+port number to the same internal port nubmer. This means that if you have several nodes running
+behind a NAT, they should each use a different RPC port number.
+
+#### `rpc_public_addr`
+
+The address and port that other nodes need to use to contact this node for
+RPC calls.  **This parameter is optional but recommended.** In case you have
+a NAT that binds the RPC port to a port that is different on your public IP,
+this field might help making it work.
+
+#### `bootstrap_peers`
+
+A list of peer identifiers on which to contact other Garage peers of this cluster.
+These peer identifiers have the following syntax:
+
+```
+<node public key>@<node public IP or hostname>:<port>
+```
+
+In the case where `rpc_public_addr` is correctly specified in the
+configuration file, the full identifier of a node including IP and port can
+be obtained by running `garage node-id` and then included directly in the
+`bootstrap_peers` list of other nodes.  Otherwise, only the node's public
+key will be returned by `garage node-id` and you will have to add the IP
+yourself.
+
+#### `consul_host` and `consul_service_name`
+
+Garage supports discovering other nodes of the cluster using Consul.
+This works only when nodes are announced in Consul by an orchestrator such as Nomad,
+as Garage is not able to announce itself.
+
+The `consul_host` parameter should be set to the hostname of the Consul server,
+and `consul_service_name` should be set to the service name under which Garage's
+RPC ports are announced.
+
+#### `sled_cache_capacity`
+
+This parameter can be used to tune the capacity of the cache used by
+[sled](https://sled.rs), the database Garage uses internally to store metadata.
+Tune this to fit the RAM you wish to make available to your Garage instance.
+More cache means faster Garage, but the default value (128MB) should be plenty
+for most use cases.
+
+#### `sled_flush_every_ms`
+
+This parameters can be used to tune the flushing interval of sled.
+Increase this if sled is thrashing your SSD, at the risk of losing more data in case
+of a power outage (though this should not matter much as data is replicated on other
+nodes). The default value, 2000ms, should be appropriate for most use cases.
+
+
+## The `[s3_api]` section
+
+#### `api_bind_addr`
+
+The IP and port on which to bind for accepting S3 API calls.
+This endpoint does not suport TLS: a reverse proxy should be used to provide it.
+
+#### `s3_region`
+
+Garage will accept S3 API calls that are targetted to the S3 region defined here.
+API calls targetted to other regions will fail with a AuthorizationHeaderMalformed error
+message that redirects the client to the correct region.
+
+
+## The `[s3_web]` section
+
+Garage allows to publish content of buckets as websites. This section configures the
+behaviour of this module.
+
+#### `bind_addr`
+
+The IP and port on which to bind for accepting HTTP requests to buckets configured
+for website access.
+This endpoint does not suport TLS: a reverse proxy should be used to provide it.
+
+#### `root_domain`
+
+The optionnal suffix appended to bucket names for the corresponding HTTP Host.
+
+For instance, if `root_domain` is `web.garage.eu`, a bucket called `deuxfleurs.fr`
+will be accessible either with hostname `deuxfleurs.fr.web.garage.eu`
+or with hostname `deuxfleurs.fr`.
+
+#### `index`
+
+The name of the index file to return for requests ending with `/` (usually `index.html`).

doc/book/src/reference_manual/s3_compatibility.md

@@ -1,6 +1,6 @@
-## S3 Compatibility status
+# S3 Compatibility status
 
-### Global S3 features
+## Global S3 features
 
 Implemented:
 
@@ -14,71 +14,44 @@ Not implemented:
 
 - vhost-style URLs (`bucket.garage.tld/key`)
 - object-level ACL
+- object versioning
 - encryption
 - most `x-amz-` headers
 
 
-### Endpoint implementation
+## Endpoint implementation
 
 All APIs that are not mentionned are not implemented and will return a 400 bad request.
 
-#### AbortMultipartUpload
+| Endpoint                     | Status                           |
+|------------------------------|----------------------------------|
+| AbortMultipartUpload         | Implemented                      |
+| CompleteMultipartUpload      | Implemented                      |
+| CopyObject                   | Implemented                      |
+| CreateBucket                 | Unsupported, stub (see below)    |
+| CreateMultipartUpload        | Implemented                      |
+| DeleteBucket                 | Unsupported (see below)          |
+| DeleteObject                 | Implemented                      |
+| DeleteObjects                | Implemented                      |
+| GetBucketLocation            | Implemented                      |
+| GetBucketVersioning          | Stub (see below)                 |
+| GetObject                    | Implemented                      |
+| HeadBucket                   | Implemented                      |
+| HeadObject                   | Implemented                      |
+| ListBuckets                  | Implemented                      |
+| ListObjects                  | Implemented, bugs? (see below)   |
+| ListObjectsV2                | Implemented                      |
+| PutObject                    | Implemented                      |
+| UploadPart                   | Implemented                      |
 
-Implemented.
 
-#### CompleteMultipartUpload
 
-Implemented badly. Garage will not check that all the parts stored correspond to the list given by the client in the request body. This means that the multipart upload might be completed with an invalid size. This is a bug and will be fixed.
+- **CreateBucket:** Garage does not yet accept creating buckets or giving access using API calls, it has to be done using the CLI tools. CreateBucket will return a 200 if the bucket exists and user has write access, and a 403 Forbidden in all other cases.
 
-#### CopyObject
+- **DeleteBucket:** Garage does not yet accept deleting buckets using API calls, it has to be done using the CLI tools. This request will return a 403 Forbidden.
 
-Implemented.
+- **GetBucketVersioning:** Stub implementation (Garage does not yet support versionning so this always returns
+"versionning not enabled").
 
-#### CreateBucket
-
-Garage does not accept creating buckets or giving access using API calls, it has to be done using the CLI tools. CreateBucket will return a 200 if the bucket exists and user has write access, and a 403 Forbidden in all other cases.
-
-#### CreateMultipartUpload
-
-Implemented.
-
-#### DeleteBucket
-
-Garage does not accept deleting buckets using API calls, it has to be done using the CLI tools. This request will return a 403 Forbidden.
-
-#### DeleteObject
-
-Implemented.
-
-#### DeleteObjects
-
-Implemented.
-
-#### GetObject
-
-Implemented.
-
-#### HeadBucket
-
-Implemented.
-
-#### HeadObject
-
-Implemented.
-
-#### ListObjects
-
-Implemented, but there isn't a very good specification of what `encoding-type=url` covers so there might be some encoding bugs. In our implementation the url-encoded fields are in the same in ListObjects as they are in ListObjectsV2.
-
-#### ListObjectsV2
-
-Implemented.
-
-#### PutObject
-
-Implemented.
-
-#### UploadPart
-
-Implemented.
+- **ListObjects:** Implemented, but there isn't a very good specification of what `encoding-type=url` covers so there might be some encoding bugs. In our implementation the url-encoded fields are in the same in ListObjects as they are in ListObjectsV2.
 

doc/book/src/working_documents/load_balancing.md

@@ -1,8 +1,8 @@
-## Load Balancing Data (planned for version 0.2)
+# Load Balancing Data (planned for version 0.2)
 
 I have conducted a quick study of different methods to load-balance data over different Garage nodes using consistent hashing.
 
-### Requirements
+## Requirements
 
 - *good balancing*: two nodes that have the same announced capacity should receive close to the same number of items
 
@@ -15,9 +15,9 @@ I have conducted a quick study of different methods to load-balance data over di
   replicas, independently of the order in which nodes were added/removed (this
   is to keep the implementation simple)
 
-### Methods
+## Methods
 
-#### Naive multi-DC ring walking strategy
+### Naive multi-DC ring walking strategy
 
 This strategy can be used with any ring-like algorithm to make it aware of the *multi-datacenter* requirement:
 
@@ -38,7 +38,7 @@ This method was implemented in the first version of Garage, with the basic
 ring construction from Dynamo DB that consists in associating `n_token` random positions to
 each node (I know it's not optimal, the Dynamo paper already studies this).
 
-#### Better rings
+### Better rings
 
 The ring construction that selects `n_token` random positions for each nodes gives a ring of positions that
 is not well-balanced: the space between the tokens varies a lot, and some partitions are thus bigger than others.
@@ -150,7 +150,7 @@ removing grisou gipsie : 49.22% 36.52% 12.79% 1.46%
 on average:  62.94% 27.89% 8.61% 0.57%                  <-- WORSE THAN PREVIOUSLY
 ```
 
-#### The magical solution: multi-DC aware MagLev
+### The magical solution: multi-DC aware MagLev
 
 Suppose we want to select three replicas for each partition (this is what we do in our simulation and in most Garage deployments).
 We apply MagLev three times consecutively, one for each replica selection.

doc/book/src/working_documents/migration_04.md

@@ -0,0 +1,105 @@
+# Migrating from 0.3 to 0.4
+
+**Migrating from 0.3 to 0.4 is unsupported. This document is only intended to
+document the process internally for the Deuxfleurs cluster where we have to do
+it. Do not try it yourself, you will lose your data and we will not help you.**
+
+**Migrating from 0.2 to 0.4 will break everything for sure. Never try it.**
+
+The internal data format of Garage hasn't changed much between 0.3 and 0.4.
+The Sled database is still the same, and the data directory as well.
+
+The following has changed, all in the meta directory:
+
+- `node_id` in 0.3 contains the identifier of the current node. In 0.4, this
+  file does nothing and should be deleted. It is replaced by `node_key` (the
+  secret key) and `node_key.pub` (the associated public key). A node's
+  identifier on the ring is its public key.
+
+- `peer_info` in 0.3 contains the list of peers saved automatically by Garage.
+  The format has changed and it is now stored in `peer_list` (`peer_info`
+  should be deleted).
+
+When migrating, all node identifiers will change. This also means that the
+affectation of data partitions on the ring will change, and lots of data will
+have to be rebalanced.
+
+- If your cluster has only 3 nodes, all nodes store everything, therefore nothing has to be rebalanced.
+
+- If your cluster has only 4 nodes, for any partition there will always be at
+  least 2 nodes that stored data before that still store it after. Therefore
+  the migration should in theory be transparent and Garage should continue to
+  work during the rebalance.
+
+- If your cluster has 5 or more nodes, data will disappear during the
+  migration. Do not migrate (fortunately we don't have this scenario at
+  Deuxfleurs), or if you do, make Garage unavailable until things stabilize
+  (disable web and api access).
+
+
+The migration steps are as follows:
+
+1. Prepare a new configuration file for 0.4. For each node, point to the same
+   meta and data directories as Garage 0.3. Basically, the things that change
+   are the following:
+
+  - No more `rpc_tls` section
+  - You have to generate a shared `rpc_secret` and put it in all config files
+  - `bootstrap_peers` has a different syntax as it has to contain node keys.
+    Leave it empty and use `garage node-id` and `garage node connect` instead (new features of 0.4)
+  - put the publicly accessible RPC address of your node in `rpc_public_addr` if possible (its optional but recommended)
+  - If you are using Consul, change the `consul_service_name` to NOT be the name advertised by Nomad.
+    Now Garage is responsible for advertising its own service itself.
+
+2. Disable api and web access for some time (Garage does not support disabling
+   these endpoints but you can change the port number or stop your reverse
+   proxy for instance).
+
+3. Do `garage repair -a --yes tables` and `garage repair -a --yes blocks`,
+   check the logs and check that all data seems to be synced correctly between
+   nodes.
+
+4. Save somewhere the output of `garage status`. We will need this to remember
+   how to reconfigure nodes in 0.4.
+
+5. Turn off Garage 0.3
+
+6. Backup metadata folders if you can (i.e. if you have space to do it
+   somewhere). Backuping data folders could also be usefull but that's much
+   harder to do. If your filesystem supports snapshots, this could be a good
+   time to use them.
+
+7. Turn on Garage 0.4
+
+8. At this point, running `garage status` should indicate that all nodes of the
+   previous cluster are "unavailable". The nodes have new identifiers that
+   should appear in healthy nodes once they can talk to one another (use
+   `garage node connect` if necessary`). They should have NO ROLE ASSIGNED at
+   the moment.
+
+9. Prepare a script with several `garage node configure` commands that replace
+   each of the v0.3 node ID with the corresponding v0.4 node ID, with the same
+   zone/tag/capacity. For example if your node `drosera` had identifier `c24e`
+   before and now has identifier `789a`, and it was configured with capacity
+   `2` in zone `dc1`, put the following command in your script:
+
+```bash
+garage node configure 789a -z dc1 -c 2 -t drosera --replace c24e
+```
+
+10. Run your reconfiguration script. Check that the new output of `garage
+	status` contains the correct node IDs with the correct values for capacity
+	and zone. Old nodes should no longer be mentioned.
+
+11. If your cluster has 4 nodes or less, and you are feeling adventurous, you
+	can reenable Web and API access now. Things will probably work.
+
+12. Garage might already be resyncing stuff. Issue a `garage repair -a --yes
+	tables` and `garage repair -a --yes blocks` to force it to do so.
+
+13. Wait for resyncing activity to stop in the logs. Do steps 12 and 13 two or
+	three times, until you see that when you issue the repair commands, nothing
+	gets resynced any longer.
+
+14. Your upgraded cluster should be in a working state. Re-enable API and Web
+	access and check that everything went well.

doc/talks/2021-09-13-ngi-kickoff/.gitignore

@@ -0,0 +1,12 @@
+*
+
+!img
+
+!.gitignore
+!*.svg
+!*.png
+!*.jpg
+!*.tex
+!Makefile
+!.gitignore
+!talk.pdf

doc/talks/2021-09-13-ngi-kickoff/Makefile

@@ -0,0 +1,3 @@
+talk.pdf: talk.tex
+	pdflatex talk.tex
+

doc/talks/2021-09-13-ngi-kickoff/talk.tex

@@ -0,0 +1,147 @@
+%\nonstopmode
+\documentclass[aspectratio=169]{beamer}
+\usepackage[utf8]{inputenc}
+% \usepackage[frenchb]{babel}
+\usepackage{amsmath}
+\usepackage{mathtools}
+\usepackage{breqn}
+\usepackage{multirow}
+\usetheme{boxes}
+\usepackage{graphicx}
+%\useoutertheme[footline=authortitle,subsection=false]{miniframes}
+
+\beamertemplatenavigationsymbolsempty
+
+\usepackage{tabu}
+\usepackage{multicol}
+\usepackage{vwcol}
+\usepackage{stmaryrd}
+\usepackage{graphicx}
+
+\usepackage[normalem]{ulem}
+
+\title{Presentation of the Garage project}
+\subtitle{NGI pointer kickoff meeting}
+\author{Deuxfleurs Association}
+\date{2021-09-13}
+
+\begin{document}
+
+\begin{frame}
+	\centering
+	\includegraphics[width=.3\linewidth]{../../sticker/Garage.pdf}
+	\vspace{1em}
+
+	{\large\bf Deuxfleurs Association}
+	\vspace{1em}
+
+	\url{https://deuxfleurs.fr/}
+
+	\url{https://garagehq.deuxfleurs.fr/}
+
+	Matrix channel: \texttt{\#garage:deuxfleurs.fr}
+\end{frame}
+
+\begin{frame}
+	\frametitle{Our objective at Deuxfleurs}
+	
+	\begin{center}
+		\textbf{Promote self-hosting and small-scale hosting\\
+			as an alternative to large cloud providers}
+	\end{center}
+	\vspace{2em}
+	\visible<2->{
+		Why is it hard?
+	}
+	\visible<3->{
+		\vspace{2em}
+		\begin{center}
+			\textbf{\underline{Resilience}}\\
+			{\footnotesize (we want good uptime/availability with low supervision)}
+		\end{center}
+	}
+\end{frame}
+
+\begin{frame}
+	\frametitle{How to be resilient (the hard way)}
+
+	Entreprise-grade systems typically employ:
+	\vspace{1em}
+	\begin{itemize}
+		\item Redundant Internet connections
+		\item Redundant electricity
+		\item UPSes
+		\item RAID
+		\item ... 
+	\end{itemize}
+	\vspace{1em}
+	$\to$ it's costly and only worth it at DC scale
+\end{frame}
+
+\begin{frame}
+	\frametitle{How to be resilient (the \underline{\textbf{cheap}} way)}
+
+	Instead, we use:
+	\vspace{1em}
+	\begin{itemize}
+		\item Commodity hardware (e.g. old desktop PCs)
+			\vspace{.5em}
+		\item<2-> Commodity Internet (e.g. FTTB, FTTH) and electricity
+			\vspace{.5em}
+		\item<3-> \textbf{Geographical redundancy} (multi-site replication)
+			\vspace{.5em}
+		\item<4-> \textbf{Fault-tolerant distributed algorithms}
+	\end{itemize}
+	\vspace{1em}
+	\visible<5->{
+		\centering
+		\underline{\textbf{This is how we build Garage.}}
+	}
+\end{frame}
+
+\begin{frame}
+	\frametitle{But what is Garage, exactly?}
+	\textbf{Garage is a self-hosted drop-in replacement for the Amazon S3 object store}\\
+	\vspace{.5em}
+	that implements resilience through geographical redundancy on commodity hardware
+
+	\vspace{1em}
+	\visible<2->{
+		\begin{center}
+				Current status: technical preview\\
+				Our goal: release a stable v1.0
+		\end{center}
+	}
+
+	\vspace{1em}
+	\visible<3->{
+		\textbf{Comming up next: an e-mail IMAP inbox based on the same principles}
+	}
+
+	\vspace{1em}
+	\visible<4->{
+		\begin{center}
+				Current status: just an idea\\
+				Our goal: at least a PoC, maybe more
+		\end{center}
+	}
+\end{frame}
+
+\begin{frame}
+	\centering
+	\includegraphics[width=.3\linewidth]{../../sticker/Garage.pdf}
+	\vspace{1em}
+
+	{\large\bf Deuxfleurs Association}
+	\vspace{1em}
+
+	\url{https://deuxfleurs.fr/}
+
+	\url{https://garagehq.deuxfleurs.fr/}
+
+	Matrix channel: \texttt{\#garage:deuxfleurs.fr}
+\end{frame}
+
+\end{document}
+
+%% vim: set ts=4 sw=4 tw=0 noet spelllang=fr :

genkeys.sh

@@ -11,7 +11,7 @@ cd pki
 # the RPC protocol will use to authenticate the other side.
 if [ ! -f garage-ca.key ]; then
     echo "Generating Garage CA keys..."
-	openssl genrsa -out garage-ca.key 4096
+	openssl genpkey -algorithm ED25519 -out garage-ca.key
 	openssl req -x509 -new -nodes -key garage-ca.key -sha256 -days 3650 -out garage-ca.crt -subj "/C=FR/O=Garage"
 fi
 
@@ -22,7 +22,7 @@ fi
 if [ ! -f garage.crt ]; then
 	echo "Generating Garage agent keys..."
 	if [ ! -f garage.key ]; then
-		openssl genrsa -out garage.key 4096
+		openssl genpkey -algorithm ED25519 -out garage.key
 	fi
 	openssl req -new -sha256 -key garage.key -subj "/C=FR/O=Garage/CN=garage" \
 		-out garage.csr
@@ -56,7 +56,7 @@ fi
 if [ ! -f garage-client.crt ]; then
 	echo "Generating Garage client keys..."
 	if [ ! -f garage-client.key ]; then
-		openssl genrsa -out garage-client.key 4096
+		openssl genpkey -algorithm ED25519 -out garage-client.key
 	fi
 	openssl req -new -sha256 -key garage-client.key -subj "/C=FR/O=Garage" \
 		-out garage-client.csr

make_docker.sh

@@ -1,20 +0,0 @@
-#!/bin/sh
-
-BIN=target/release/garage
-DOCKER=lxpz/garage_amd64
-
-TAG=$1
-if [ -z "$1" ]; then
-	echo "Usage: $0 <tag>"
-	exit 1
-fi
-
-RUSTFLAGS="-C link-arg=-fuse-ld=lld -C target-cpu=x86-64 -C target-feature=+sse2" cargo build --release --no-default-features
-cp $BIN $BIN.stripped
-strip $BIN.stripped
-
-docker pull archlinux:latest
-docker build -t $DOCKER:$TAG .
-docker push $DOCKER:$TAG
-docker tag $DOCKER:$TAG $DOCKER:latest
-docker push $DOCKER:latest

nix/build_index.nix

@@ -0,0 +1,146 @@
+{
+  path ? "/../aws-list.txt",
+}:
+
+with import ./common.nix;
+let 
+  pkgs = import pkgsSrc {};
+  lib = pkgs.lib;
+
+  /* Converts a key list and a value list to a set
+
+     Example:
+       listToSet [ "name" "version" ] [ "latex" "3.14" ]
+       => { name = "latex"; version = "3.14"; }
+  */
+  listToSet = keys: values: 
+    builtins.listToAttrs 
+      (lib.zipListsWith 
+        (a: b: { name = a; value = b; }) 
+        keys
+        values);
+
+  /* Says if datetime a is more recent than datetime b
+
+    Example:
+      cmpDate { date = "2021-09-10"; time = "22:12:15"; } { date = "2021-02-03"; time = "23:54:12"; }
+      => true
+  */
+  cmpDate = a: b: 
+    let da = (builtins.head a.builds).date;
+        db = (builtins.head b.builds).date;
+    in 
+      if da == db then  (builtins.head a.builds).time > (builtins.head b.builds).time
+      else da > db;
+
+  /* Pretty platforms */
+  prettyPlatform = name:
+    if name == "aarch64-unknown-linux-musl" then "linux/arm64"
+    else if name == "armv6l-unknown-linux-musleabihf" then "linux/arm"
+    else if name == "x86_64-unknown-linux-musl" then "linux/amd64"
+    else if name == "i686-unknown-linux-musl" then "linux/386"
+    else name;
+
+  /* Parsing */
+  list = builtins.readFile (./. + path);
+  entries = lib.splitString "\n" list;
+
+  elems = builtins.filter 
+    (e: (builtins.length e) == 4) 
+    (map 
+      (x: builtins.filter (e: e != "") (lib.splitString " " x)) 
+      entries);
+
+  keys = ["date" "time" "size" "path"];
+  parsed = map (entry: listToSet keys entry) elems;
+
+  subkeys = ["root" "version" "platform" "binary" ];
+  builds = map (entry: entry // listToSet subkeys (lib.splitString "/" entry.path)) parsed;
+
+  /* Aggregation */
+  builds_per_version = lib.foldl (acc: v: acc // { ${v.version} = if builtins.hasAttr v.version acc then acc.${v.version} ++ [ v ] else [ v ]; }) {} builds; 
+
+  versions = builtins.attrNames builds_per_version;
+  versions_release = builtins.filter (x: builtins.match "v[0-9]+\.[0-9]+\.[0-9]+(\.[0-9]+)?" x != null) versions;
+  versions_commit = builtins.filter (x: builtins.match "[0-9a-f]{40}" x != null) versions;
+  versions_extra = lib.subtractLists (versions_release ++ versions_commit) versions;
+
+  sorted_builds = [
+    {
+      name = "Release";
+      hide = false;
+      type = "tag";
+      description = "Release builds are the official builds, they are tailored for productions and are the most tested.";
+      builds =  builtins.sort (a: b: a.version > b.version) (map (x: { version = x; builds = builtins.getAttr x builds_per_version; }) versions_release);
+    }
+    {
+      name = "Extra";
+      hide = true;
+      type = "tag";
+      description = "Extra builds are built on demand to test a specific feature or a specific need.";
+      builds = builtins.sort cmpDate (map (x: { version = x; builds = builtins.getAttr x builds_per_version; }) versions_extra);
+    }
+    {
+      name = "Development";
+      hide = true;
+      type = "commit";
+      description = "Development builds are built periodically. Use them if you want to test a specific feature that is not yet released.";
+      builds = builtins.sort cmpDate (map (x: { version = x; builds = builtins.getAttr x builds_per_version; }) versions_commit);
+    }
+  ];
+
+in
+  pkgs.writeText "index.html" ''
+<!doctype html>
+<html>
+  <head>
+    <meta charset="utf-8" />
+    <title>Garage releases</title>
+    <style>
+      html, body { margin:0; padding: 0 }
+      body { font-family: 'Helvetica', Sans; }
+      section { margin: 1rem; }
+      ul { padding:0; margin: 0.2rem }
+      li {
+        border-radius: 0.2rem;
+        display: inline; 
+        border: 2px #0b5d83 solid; 
+        padding: 0.5rem; 
+        line-height: 3rem;
+        color: #0b5d83;
+      }
+      li:hover { background-color: #0b5d83; color: #fff; }
+      li a, li a:hover { color: inherit; text-decoration: none }
+    </style>
+  </head>
+  <body>
+   ${ builtins.toString (lib.forEach sorted_builds (r: ''
+    <section>
+      <h2>${r.name} builds</h2>
+
+      <p>${r.description}</p>
+
+      ${if r.hide then "<details><summary>Show ${r.name} builds</summary>" else ""}
+      ${ builtins.toString (lib.forEach r.builds (x: ''
+        <h3> ${x.version} (${(builtins.head x.builds).date}) </h3>
+        <p>See this build on</p>
+        <p> Binaries:
+        <ul>
+        ${ builtins.toString (lib.forEach x.builds (b: ''
+          <li><a href="/${b.path}">${prettyPlatform b.platform}</a></li>
+        ''))}
+        </ul></p>
+        <p> Sources:
+        <ul>
+         <li><a href="https://git.deuxfleurs.fr/Deuxfleurs/garage/src/${r.type}/${x.version}">gitea</a></li>
+         <li><a href="https://git.deuxfleurs.fr/Deuxfleurs/garage/archive/${x.version}.zip">.zip</a></li>
+         <li><a href="https://git.deuxfleurs.fr/Deuxfleurs/garage/archive/${x.version}.tar.gz">.tar.gz</a></li>
+        </ul></p>
+      ''))  }
+      ${ if builtins.length r.builds == 0 then "<em>There is no build for this category</em>" else "" }
+      ${if r.hide then "</details>" else ""}
+    </section>
+  ''))}
+  </body>
+</html>
+''

nix/common.nix

@@ -0,0 +1,21 @@
+rec {
+  /*
+   * Fixed dependencies
+   */
+  pkgsSrc = fetchTarball {
+    # As of 2021-10-04
+    url ="https://github.com/NixOS/nixpkgs/archive/b27d18a412b071f5d7991d1648cfe78ee7afe68a.tar.gz";
+    sha256 = "1xy9zpypqfxs5gcq5dcla4bfkhxmh5nzn9dyqkr03lqycm9wg5cr";
+  };
+  cargo2nixSrc = fetchGit {
+    # As of 2021-10-06
+    url = "https://github.com/superboum/cargo2nix";
+    rev = "1364752cd784764db2ef5b1e1248727cebfae2ce";
+  };
+
+  /*
+   * Shared objects
+   */
+  cargo2nix = import cargo2nixSrc;
+  cargo2nixOverlay = import "${cargo2nixSrc}/overlay";
+}

nix/kaniko.nix

@@ -0,0 +1,23 @@
+pkgs:
+pkgs.buildGoModule rec {
+  pname = "kaniko";
+  version = "1.6.0";
+
+  src = pkgs.fetchFromGitHub {
+    owner = "GoogleContainerTools";
+    repo = "kaniko";
+    rev = "v${version}";
+    sha256 = "1fnclr556avxay6pvgw5ya3xbxfnf2gv4njq2hr4fd6fcjyslq5h";
+  };
+
+  vendorSha256 = null;
+
+  checkPhase = "true";
+
+  meta = with pkgs.lib; {
+    description = "kaniko is a tool to build container images from a Dockerfile, inside a container or Kubernetes cluster.";
+    homepage = "https://github.com/GoogleContainerTools/kaniko";
+    license = licenses.asl20;
+    platforms = platforms.linux;
+  };
+}

nix/nix.conf

@@ -0,0 +1,4 @@
+substituters = https://cache.nixos.org https://nix.web.deuxfleurs.fr
+trusted-public-keys = cache.nixos.org-1:6NCHdD59X431o0gWypbMrAURkbJ16ZPMQFGspcDShjY= nix.web.deuxfleurs.fr:eTGL6kvaQn6cDR/F9lDYUIP9nCVR/kkshYfLDJf1yKs=
+max-jobs = auto
+cores = 4

nix/toolchain.nix

@@ -0,0 +1,29 @@
+{
+  system ? builtins.currentSystem,
+}:
+
+with import ./common.nix;
+
+let
+  platforms = [
+    "x86_64-unknown-linux-musl"
+    "i686-unknown-linux-musl"
+    "aarch64-unknown-linux-musl"
+    "armv6l-unknown-linux-musleabihf"
+  ];
+  pkgsList = builtins.map (target: import pkgsSrc { 
+    inherit system; 
+    crossSystem = { config = target; };
+  }) platforms;
+  pkgsHost = import pkgsSrc {};
+  lib = pkgsHost.lib;
+  kaniko = (import ./kaniko.nix) pkgsHost;
+in 
+  lib.flatten (builtins.map (pkgs: [
+     pkgs.rustPlatform.rust.rustc
+     pkgs.rustPlatform.rust.cargo
+     pkgs.buildPackages.stdenv.cc
+  ]) pkgsList) ++ [
+    kaniko
+  ]
+

script/builder_image/Dockerfile

@@ -1,7 +0,0 @@
-FROM rust:buster
-RUN apt-get update && \
-	apt-get install --yes libsodium-dev awscli python-pip wget rclone openssl socat && \
-	    rm -rf /var/lib/apt/lists/*
-RUN wget https://dl.min.io/client/mc/release/linux-amd64/mc -O /usr/local/bin/mc && chmod +x /usr/local/bin/mc
-RUN rustup component add rustfmt clippy
-RUN pip install s3cmd

script/builder_image/Makefile

@@ -1,8 +0,0 @@
-DOCKER=lxpz/garage_builder_amd64
-
-docker:
-	docker build -t $(DOCKER):$(TAG) .
-	docker push $(DOCKER):$(TAG)
-	docker tag $(DOCKER):$(TAG) $(DOCKER):latest
-	docker push $(DOCKER):latest
-	

script/dev-bucket.sh

@@ -6,13 +6,14 @@ SCRIPT_FOLDER="`dirname \"$0\"`"
 REPO_FOLDER="${SCRIPT_FOLDER}/../"
 GARAGE_DEBUG="${REPO_FOLDER}/target/debug/"
 GARAGE_RELEASE="${REPO_FOLDER}/target/release/"
-PATH="${GARAGE_DEBUG}:${GARAGE_RELEASE}:$PATH"
+NIX_RELEASE="${REPO_FOLDER}/result/bin/"
+PATH="${GARAGE_DEBUG}:${GARAGE_RELEASE}:${NIX_RELEASE}:$PATH"
 
-garage bucket create eprouvette
-KEY_INFO=`garage key new --name opérateur`
+garage -c /tmp/config.1.toml bucket create eprouvette
+KEY_INFO=$(garage -c /tmp/config.1.toml key new --name opérateur)
 ACCESS_KEY=`echo $KEY_INFO|grep -Po 'GK[a-f0-9]+'`
 SECRET_KEY=`echo $KEY_INFO|grep -Po 'Secret key: [a-f0-9]+'|grep -Po '[a-f0-9]+$'`
-garage bucket allow eprouvette --read --write --key $ACCESS_KEY
+garage -c /tmp/config.1.toml bucket allow eprouvette --read --write --key $ACCESS_KEY
 echo "$ACCESS_KEY $SECRET_KEY" > /tmp/garage.s3
 
 echo "Bucket s3://eprouvette created. Credentials stored in /tmp/garage.s3."

script/dev-cluster.sh

@@ -6,7 +6,8 @@ SCRIPT_FOLDER="`dirname \"$0\"`"
 REPO_FOLDER="${SCRIPT_FOLDER}/../"
 GARAGE_DEBUG="${REPO_FOLDER}/target/debug/"
 GARAGE_RELEASE="${REPO_FOLDER}/target/release/"
-PATH="${GARAGE_DEBUG}:${GARAGE_RELEASE}:$PATH"
+NIX_RELEASE="${REPO_FOLDER}/result/bin/"
+PATH="${GARAGE_DEBUG}:${GARAGE_RELEASE}:${NIX_RELEASE}:$PATH"
 FANCYCOLORS=("41m" "42m" "44m" "45m" "100m" "104m")
 
 export RUST_BACKTRACE=1 
@@ -16,6 +17,10 @@ MAIN_LABEL="\e[${FANCYCOLORS[0]}[main]\e[49m"
 WHICH_GARAGE=$(which garage || exit 1)
 echo -en "${MAIN_LABEL} Found garage at: ${WHICH_GARAGE}\n"
 
+NETWORK_SECRET="$(openssl rand -hex 32)"
+
+
+# <<<<<<<<< BEGIN FOR LOOP ON NODES
 for count in $(seq 1 3); do
 CONF_PATH="/tmp/config.$count.toml"
 LABEL="\e[${FANCYCOLORS[$count]}[$count]\e[49m"
@@ -25,13 +30,10 @@ block_size = 1048576			# objects are split in blocks of maximum this number of b
 metadata_dir = "/tmp/garage-meta-$count"
 data_dir = "/tmp/garage-data-$count"
 rpc_bind_addr = "0.0.0.0:$((3900+$count))"		# the port other Garage nodes will use to talk to this node
-bootstrap_peers = [
-  "127.0.0.1:3901",
-  "127.0.0.1:3902",
-  "127.0.0.1:3903"
-]
-max_concurrent_rpc_requests = 12
+rpc_public_addr = "127.0.0.1:$((3900+$count))"
+bootstrap_peers = []
 replication_mode = "3"
+rpc_secret = "$NETWORK_SECRET"
 
 [s3_api]
 api_bind_addr = "0.0.0.0:$((3910+$count))"	# the S3 API port, HTTP without TLS. Add a reverse proxy for the TLS part.
@@ -60,13 +62,30 @@ if [ -z "$SKIP_HTTPS" ]; then
   socat openssl-listen:4443,reuseaddr,fork,cert=/tmp/garagessl/test.pem,verify=0 tcp4-connect:localhost:3911 &
 fi
 
-(garage server -c /tmp/config.$count.toml 2>&1|while read r; do echo -en "$LABEL $r\n"; done) &
+(garage -c /tmp/config.$count.toml server 2>&1|while read r; do echo -en "$LABEL $r\n"; done) &
+done
+# >>>>>>>>>>>>>>>> END FOR LOOP ON NODES
+
+sleep 3
+# Establish connections between nodes
+for count in $(seq 1 3); do
+	NODE=$(garage -c /tmp/config.$count.toml node-id -q)
+	for count2 in $(seq 1 3); do
+		garage -c /tmp/config.$count2.toml node connect $NODE
+	done
 done
 
-until garage status 2>&1|grep -q Healthy ; do 
+RETRY=120
+until garage -c /tmp/config.1.toml status 2>&1|grep -q HEALTHY ; do 
+  (( RETRY-- ))
+  if (( RETRY <= 0 )); then 
+    echo -en "${MAIN_LABEL} Garage did not start"
+    exit 1
+  fi
 	echo -en "${MAIN_LABEL} cluster starting...\n"
 	sleep 1
 done
+
 echo -en "${MAIN_LABEL} cluster started\n"
 
 wait

script/dev-configure.sh

@@ -6,18 +6,25 @@ SCRIPT_FOLDER="`dirname \"$0\"`"
 REPO_FOLDER="${SCRIPT_FOLDER}/../"
 GARAGE_DEBUG="${REPO_FOLDER}/target/debug/"
 GARAGE_RELEASE="${REPO_FOLDER}/target/release/"
-PATH="${GARAGE_DEBUG}:${GARAGE_RELEASE}:$PATH"
+NIX_RELEASE="${REPO_FOLDER}/result/bin/"
+PATH="${GARAGE_DEBUG}:${GARAGE_RELEASE}:${NIX_RELEASE}:$PATH"
 
 sleep 5
-until garage status 2>&1|grep -q Healthy ; do 
+RETRY=120
+until garage -c /tmp/config.1.toml status 2>&1|grep -q HEALTHY ; do 
+  (( RETRY-- ))
+  if (( RETRY <= 0 )); then
+    echo "garage did not start in time, failing."
+    exit 1
+  fi
 	echo "cluster starting..."
 	sleep 1
 done
 
-garage status \
-	| grep UNCONFIGURED \
+garage -c /tmp/config.1.toml status \
+	| grep 'NO ROLE' \
 	| grep -Po '^[0-9a-f]+' \
 	| while read id; do 
-	  garage node configure -z dc1 -c 1 $id
+	  garage -c /tmp/config.1.toml node configure -z dc1 -c 1 $id
 	done
 

script/test-smoke.sh

@@ -8,22 +8,23 @@ SCRIPT_FOLDER="`dirname \"$0\"`"
 REPO_FOLDER="${SCRIPT_FOLDER}/../"
 GARAGE_DEBUG="${REPO_FOLDER}/target/debug/"
 GARAGE_RELEASE="${REPO_FOLDER}/target/release/"
-PATH="${GARAGE_DEBUG}:${GARAGE_RELEASE}:$PATH"
+NIX_RELEASE="${REPO_FOLDER}/result/bin/"
+PATH="${GARAGE_DEBUG}:${GARAGE_RELEASE}:${NIX_RELEASE}:$PATH"
 
 # @FIXME Duck is not ready for testing, we have a bug
 SKIP_DUCK=1
 
 echo "⏳ Setup"
-cargo build
 ${SCRIPT_FOLDER}/dev-clean.sh
 ${SCRIPT_FOLDER}/dev-cluster.sh > /tmp/garage.log 2>&1 &
+sleep 6
 ${SCRIPT_FOLDER}/dev-configure.sh
 ${SCRIPT_FOLDER}/dev-bucket.sh
 
 which garage
-garage status
-garage key list
-garage bucket list
+garage -c /tmp/config.1.toml status
+garage -c /tmp/config.1.toml key list
+garage -c /tmp/config.1.toml bucket list
 
 dd if=/dev/urandom of=/tmp/garage.1.rnd bs=1k count=2 # No multipart, inline storage (< INLINE_THRESHOLD = 3072 bytes)
 dd if=/dev/urandom of=/tmp/garage.2.rnd bs=1M count=5 # No multipart but file will be chunked
@@ -116,9 +117,9 @@ if [ -z "$SKIP_AWS" ]; then
   echo "<h1>hello world</h1>" > /tmp/garage-index.html
   aws s3 cp /tmp/garage-index.html s3://eprouvette/index.html
   [ `curl -s -o /dev/null -w "%{http_code}" --header "Host: eprouvette.garage.tld"  http://127.0.0.1:3923/ ` == 404 ]
-  garage bucket website --allow eprouvette
+  garage -c /tmp/config.1.toml bucket website --allow eprouvette
   [ `curl -s -o /dev/null -w "%{http_code}" --header "Host: eprouvette.garage.tld"  http://127.0.0.1:3923/ ` == 200 ]
-  garage bucket website --deny eprouvette
+  garage -c /tmp/config.1.toml bucket website --deny eprouvette
   [ `curl -s -o /dev/null -w "%{http_code}" --header "Host: eprouvette.garage.tld"  http://127.0.0.1:3923/ ` == 404 ]
   aws s3 rm s3://eprouvette/index.html
   rm /tmp/garage-index.html
@@ -127,8 +128,8 @@ fi
 echo "🏁 Teardown"
 AWS_ACCESS_KEY_ID=`cat /tmp/garage.s3 |cut -d' ' -f1`
 AWS_SECRET_ACCESS_KEY=`cat /tmp/garage.s3 |cut -d' ' -f2`
-garage bucket deny --read --write eprouvette --key $AWS_ACCESS_KEY_ID
-garage bucket delete --yes eprouvette
-garage key delete --yes $AWS_ACCESS_KEY_ID
+garage -c /tmp/config.1.toml bucket deny --read --write eprouvette --key $AWS_ACCESS_KEY_ID
+garage -c /tmp/config.1.toml bucket delete --yes eprouvette
+garage -c /tmp/config.1.toml key delete --yes $AWS_ACCESS_KEY_ID
 
 echo "✅ Success"

shell.nix

@@ -0,0 +1,93 @@
+{
+  system ? builtins.currentSystem,
+  rust ? true,
+  integration ? true,
+  release ? true,
+}:
+
+with import ./nix/common.nix;
+
+let
+  pkgs = import pkgsSrc {
+    inherit system;
+    overlays = [ cargo2nixOverlay ];
+  };
+  kaniko = (import ./nix/kaniko.nix) pkgs;
+
+in
+
+pkgs.mkShell {
+  shellHook = ''
+function to_s3 {
+  aws \
+      --endpoint-url https://garage.deuxfleurs.fr \
+      --region garage \
+    s3 cp \
+      ./result/bin/garage \
+      s3://garagehq.deuxfleurs.fr/_releases/''${DRONE_TAG:-$DRONE_COMMIT}/''${TARGET}/garage
+}
+
+function to_docker {
+  executor  \
+    --force \
+    --customPlatform="''${DOCKER_PLATFORM}" \
+    --destination "''${CONTAINER_NAME}:''${CONTAINER_TAG}" \
+    --context dir://`pwd` \
+    --verbosity=debug
+}
+
+function refresh_index {
+  aws \
+      --endpoint-url https://garage.deuxfleurs.fr \
+      --region garage \
+    s3 ls \
+      --recursive \
+      s3://garagehq.deuxfleurs.fr/_releases/ \
+  > aws-list.txt
+
+  nix-build nix/build_index.nix
+
+  aws \
+      --endpoint-url https://garage.deuxfleurs.fr \
+      --region garage \
+    s3 cp \
+      --content-type "text/html" \
+      result \
+      s3://garagehq.deuxfleurs.fr/_releases.html
+}
+
+function refresh_toolchain {
+  nix copy \
+    --to 's3://nix?endpoint=garage.deuxfleurs.fr&region=garage&secret-key=/etc/nix/signing-key.sec' \
+    $(nix-store -qR \
+      $(nix-build --quiet --no-build-output --no-out-link nix/toolchain.nix))
+}
+  '';
+
+  nativeBuildInputs = 
+   (if rust then [
+     pkgs.rustPlatform.rust.rustc
+     pkgs.rustPlatform.rust.cargo
+     pkgs.clippy
+     pkgs.rustfmt
+     /*(pkgs.callPackage cargo2nix {}).package*/
+    ] else [])
+   ++
+   (if integration then [ 
+     pkgs.s3cmd
+     pkgs.awscli2
+     pkgs.minio-client
+     pkgs.rclone
+     pkgs.socat
+     pkgs.psmisc
+     pkgs.which
+     pkgs.openssl
+     pkgs.curl
+    ] else [])
+   ++
+   (if release then [ 
+     pkgs.awscli2
+     kaniko 
+    ] else [])
+   ;
+}

src/api/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "garage_api"
-version = "0.3.0"
+version = "0.4.0"
 authors = ["Alex Auvolat <alex@adnab.me>"]
 edition = "2018"
 license = "AGPL-3.0"
@@ -13,9 +13,9 @@ path = "lib.rs"
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [dependencies]
-garage_model = { version = "0.3.0", path = "../model" }
-garage_table = { version = "0.3.0", path = "../table" }
-garage_util = { version = "0.3.0", path = "../util" }
+garage_model = { version = "0.4.0", path = "../model" }
+garage_table = { version = "0.4.0", path = "../table" }
+garage_util = { version = "0.4.0", path = "../util" }
 
 base64 = "0.13"
 bytes = "1.0"
@@ -35,7 +35,7 @@ tokio = { version = "1.0", default-features = false, features = ["rt", "rt-multi
 http = "0.2"
 httpdate = "0.3"
 http-range = "0.1"
-hyper = "0.14"
+hyper = { version = "0.14", features = ["server", "http1", "runtime", "tcp", "stream"] }
 percent-encoding = "2.1.0"
 roxmltree = "0.14"
 serde = { version = "1.0", features = ["derive"] }

src/api/api_server.rs

@@ -39,7 +39,7 @@ pub async fn run_api_server(
 		}
 	});
 
-	let server = Server::bind(&addr).serve(service);
+	let server = Server::bind(addr).serve(service);
 
 	let graceful = server.with_graceful_shutdown(shutdown_signal);
 	info!("API server listening on http://{}", addr);
@@ -88,8 +88,8 @@ async fn handler_inner(garage: Arc<Garage>, req: Request<Body>) -> Result<Respon
 
 	let (bucket, key) = parse_bucket_key(&path)?;
 	let allowed = match req.method() {
-		&Method::HEAD | &Method::GET => api_key.allow_read(&bucket),
-		_ => api_key.allow_write(&bucket),
+		&Method::HEAD | &Method::GET => api_key.allow_read(bucket),
+		_ => api_key.allow_write(bucket),
 	};
 	if !allowed {
 		return Err(Error::Forbidden(
@@ -109,11 +109,11 @@ async fn handler_inner(garage: Arc<Garage>, req: Request<Body>) -> Result<Respon
 		match *req.method() {
 			Method::HEAD => {
 				// HeadObject query
-				Ok(handle_head(garage, &req, &bucket, &key).await?)
+				Ok(handle_head(garage, &req, bucket, key).await?)
 			}
 			Method::GET => {
 				// GetObject query
-				Ok(handle_get(garage, &req, &bucket, &key).await?)
+				Ok(handle_get(garage, &req, bucket, key).await?)
 			}
 			Method::PUT => {
 				if params.contains_key(&"partnumber".to_string())
@@ -125,8 +125,8 @@ async fn handler_inner(garage: Arc<Garage>, req: Request<Body>) -> Result<Respon
 					Ok(handle_put_part(
 						garage,
 						req,
-						&bucket,
-						&key,
+						bucket,
+						key,
 						part_number,
 						upload_id,
 						content_sha256,
@@ -136,46 +136,43 @@ async fn handler_inner(garage: Arc<Garage>, req: Request<Body>) -> Result<Respon
 					// CopyObject query
 					let copy_source = req.headers().get("x-amz-copy-source").unwrap().to_str()?;
 					let copy_source =
-						percent_encoding::percent_decode_str(&copy_source).decode_utf8()?;
+						percent_encoding::percent_decode_str(copy_source).decode_utf8()?;
 					let (source_bucket, source_key) = parse_bucket_key(&copy_source)?;
-					if !api_key.allow_read(&source_bucket) {
+					if !api_key.allow_read(source_bucket) {
 						return Err(Error::Forbidden(format!(
 							"Reading from bucket {} not allowed for this key",
 							source_bucket
 						)));
 					}
 					let source_key = source_key.ok_or_bad_request("No source key specified")?;
-					Ok(
-						handle_copy(garage, &req, &bucket, &key, &source_bucket, &source_key)
-							.await?,
-					)
+					Ok(handle_copy(garage, &req, bucket, key, source_bucket, source_key).await?)
 				} else {
 					// PutObject query
-					Ok(handle_put(garage, req, &bucket, &key, content_sha256).await?)
+					Ok(handle_put(garage, req, bucket, key, content_sha256).await?)
 				}
 			}
 			Method::DELETE => {
 				if params.contains_key(&"uploadid".to_string()) {
 					// AbortMultipartUpload query
 					let upload_id = params.get("uploadid").unwrap();
-					Ok(handle_abort_multipart_upload(garage, &bucket, &key, upload_id).await?)
+					Ok(handle_abort_multipart_upload(garage, bucket, key, upload_id).await?)
 				} else {
 					// DeleteObject query
-					Ok(handle_delete(garage, &bucket, &key).await?)
+					Ok(handle_delete(garage, bucket, key).await?)
 				}
 			}
 			Method::POST => {
 				if params.contains_key(&"uploads".to_string()) {
 					// CreateMultipartUpload call
-					Ok(handle_create_multipart_upload(garage, &req, &bucket, &key).await?)
+					Ok(handle_create_multipart_upload(garage, &req, bucket, key).await?)
 				} else if params.contains_key(&"uploadid".to_string()) {
 					// CompleteMultipartUpload call
 					let upload_id = params.get("uploadid").unwrap();
 					Ok(handle_complete_multipart_upload(
 						garage,
 						req,
-						&bucket,
-						&key,
+						bucket,
+						key,
 						upload_id,
 						content_sha256,
 					)

src/api/error.rs

@@ -82,7 +82,11 @@ impl Error {
 		match self {
 			Error::NotFound => StatusCode::NOT_FOUND,
 			Error::Forbidden(_) => StatusCode::FORBIDDEN,
-			Error::InternalError(GarageError::Rpc(_)) => StatusCode::SERVICE_UNAVAILABLE,
+			Error::InternalError(
+				GarageError::Timeout
+				| GarageError::RemoteError(_)
+				| GarageError::Quorum(_, _, _, _),
+			) => StatusCode::SERVICE_UNAVAILABLE,
 			Error::InternalError(_) | Error::Hyper(_) | Error::Http(_) => {
 				StatusCode::INTERNAL_SERVER_ERROR
 			}
@@ -95,7 +99,11 @@ impl Error {
 			Error::NotFound => "NoSuchKey",
 			Error::Forbidden(_) => "AccessDenied",
 			Error::AuthorizationHeaderMalformed(_) => "AuthorizationHeaderMalformed",
-			Error::InternalError(GarageError::Rpc(_)) => "ServiceUnavailable",
+			Error::InternalError(
+				GarageError::Timeout
+				| GarageError::RemoteError(_)
+				| GarageError::Quorum(_, _, _, _),
+			) => "ServiceUnavailable",
 			Error::InternalError(_) | Error::Hyper(_) | Error::Http(_) => "InternalError",
 			_ => "InvalidRequest",
 		}

src/api/s3_delete.rs

@@ -1,6 +1,6 @@
 use std::sync::Arc;
 
-use hyper::{Body, Request, Response};
+use hyper::{Body, Request, Response, StatusCode};
 
 use garage_util::data::*;
 use garage_util::time::*;
@@ -55,7 +55,8 @@ async fn handle_delete_internal(
 	);
 
 	garage.object_table.insert(&object).await?;
-	return Ok((deleted_version, version_uuid));
+
+	Ok((deleted_version, version_uuid))
 }
 
 pub async fn handle_delete(
@@ -68,6 +69,7 @@ pub async fn handle_delete(
 
 	Ok(Response::builder()
 		.header("x-amz-version-id", hex::encode(delete_marker_version))
+		.status(StatusCode::NO_CONTENT)
 		.body(Body::from(vec![]))
 		.unwrap())
 }
@@ -81,7 +83,7 @@ pub async fn handle_delete_objects(
 	let body = hyper::body::to_bytes(req.into_body()).await?;
 	verify_signed_content(content_sha256, &body[..])?;
 
-	let cmd_xml = roxmltree::Document::parse(&std::str::from_utf8(&body)?)?;
+	let cmd_xml = roxmltree::Document::parse(std::str::from_utf8(&body)?)?;
 	let cmd = parse_delete_objects_xml(&cmd_xml).ok_or_bad_request("Invalid delete XML query")?;
 
 	let mut ret_deleted = Vec::new();

src/api/s3_get.rs

@@ -106,12 +106,12 @@ pub async fn handle_head(
 		_ => unreachable!(),
 	};
 
-	if let Some(cached) = try_answer_cached(&version, version_meta, req) {
+	if let Some(cached) = try_answer_cached(version, version_meta, req) {
 		return Ok(cached);
 	}
 
 	let body: Body = Body::empty();
-	let response = object_headers(&version, version_meta)
+	let response = object_headers(version, version_meta)
 		.header("Content-Length", format!("{}", version_meta.size))
 		.status(StatusCode::OK)
 		.body(body)
@@ -149,7 +149,7 @@ pub async fn handle_get(
 		ObjectVersionData::FirstBlock(meta, _) => meta,
 	};
 
-	if let Some(cached) = try_answer_cached(&last_v, last_v_meta, req) {
+	if let Some(cached) = try_answer_cached(last_v, last_v_meta, req) {
 		return Ok(cached);
 	}
 
@@ -179,7 +179,7 @@ pub async fn handle_get(
 		.await;
 	}
 
-	let resp_builder = object_headers(&last_v, last_v_meta)
+	let resp_builder = object_headers(last_v, last_v_meta)
 		.header("Content-Length", format!("{}", last_v_meta.size))
 		.status(StatusCode::OK);
 
@@ -190,7 +190,7 @@ pub async fn handle_get(
 			Ok(resp_builder.body(body)?)
 		}
 		ObjectVersionData::FirstBlock(_, first_block_hash) => {
-			let read_first_block = garage.block_manager.rpc_get_block(&first_block_hash);
+			let read_first_block = garage.block_manager.rpc_get_block(first_block_hash);
 			let get_next_blocks = garage.version_table.get(&last_v.uuid, &EmptyKey);
 
 			let (first_block, version) = futures::try_join!(read_first_block, get_next_blocks)?;

src/api/s3_list.rs

@@ -68,26 +68,63 @@ pub async fn handle_list(
 	let mut result_keys = BTreeMap::<String, ListResultInfo>::new();
 	let mut result_common_prefixes = BTreeSet::<String>::new();
 
-	let mut next_chunk_start = if query.is_v2 {
+	// Determine the key from where we want to start fetch objects
+	// from the database, and whether the object at this key must
+	// be included or excluded from the response.
+	// This key can be the prefix in the base case, or intermediate
+	// points in the dataset if we are continuing a previous listing.
+	#[allow(clippy::collapsible_else_if)]
+	let (mut next_chunk_start, mut next_chunk_exclude_start) = if query.is_v2 {
 		if let Some(ct) = &query.continuation_token {
-			String::from_utf8(base64::decode(ct.as_bytes())?)?
+			// In V2 mode, the continuation token is defined as an opaque
+			// string in the spec, so we can do whatever we want with it.
+			// In our case, it is defined as either [ or ] (for include
+			// and exclude, respectively), followed by a base64 string
+			// representing the key to start with.
+			let exclude = match &ct[..1] {
+				"[" => false,
+				"]" => true,
+				_ => return Err(Error::BadRequest("Invalid continuation token".to_string())),
+			};
+			(
+				String::from_utf8(base64::decode(ct[1..].as_bytes())?)?,
+				exclude,
+			)
+		} else if let Some(sa) = &query.start_after {
+			// StartAfter has defined semantics in the spec:
+			// start listing at the first key immediately after.
+			(sa.clone(), true)
 		} else {
-			query
-				.start_after
-				.clone()
-				.unwrap_or_else(|| query.prefix.clone())
+			// In the case where neither is specified, we start
+			// listing at the specified prefix. If an object has this
+			// exact same key, we include it. (TODO is this correct?)
+			(query.prefix.clone(), false)
 		}
 	} else {
-		query.marker.clone().unwrap_or_else(|| query.prefix.clone())
+		if let Some(mk) = &query.marker {
+			// In V1 mode, the spec defines the Marker value to mean
+			// the same thing as the StartAfter value in V2 mode.
+			(mk.clone(), true)
+		} else {
+			// Base case, same as in V2 mode
+			(query.prefix.clone(), false)
+		}
 	};
 
 	debug!(
-		"List request: `{:?}` {} `{}`",
-		query.delimiter, query.max_keys, query.prefix
+		"List request: `{:?}` {} `{}`, start from {}, exclude first {}",
+		query.delimiter, query.max_keys, query.prefix, next_chunk_start, next_chunk_exclude_start
 	);
 
+	// `truncated` is a boolean that determines whether there are
+	// more items to be added.
 	let truncated;
+	// `last_processed_item` is the key of the last item
+	// that was included in the listing before truncating.
+	let mut last_processed_item = None;
+
 	'query_loop: loop {
+		// Fetch objects
 		let objects = garage
 			.object_table
 			.get_range(
@@ -103,64 +140,120 @@ pub async fn handle_list(
 			query.max_keys + 1,
 			objects.len()
 		);
+		let current_chunk_start = next_chunk_start.clone();
 
+		// Iterate on returned objects and add them to the response.
+		// If a delimiter is specified, we take care of grouping objects
+		// into CommonPrefixes.
 		for object in objects.iter() {
+			// If we have retrieved an object that doesn't start with
+			// the prefix, we know we have finished listing our stuff.
 			if !object.key.starts_with(&query.prefix) {
-				truncated = None;
+				truncated = false;
 				break 'query_loop;
 			}
 
-			if query.is_v2 && query.start_after.as_ref() == Some(&object.key) {
+			// Exclude the starting key if we have to.
+			if object.key == next_chunk_start && next_chunk_exclude_start {
 				continue;
 			}
 
-			if let Some(version) = object.versions().iter().find(|x| x.is_data()) {
-				if result_keys.len() + result_common_prefixes.len() >= query.max_keys {
-					truncated = Some(object.key.to_string());
-					break 'query_loop;
-				}
-				let common_prefix = if let Some(delimiter) = &query.delimiter {
-					let relative_key = &object.key[query.prefix.len()..];
-					relative_key
-						.find(delimiter)
-						.map(|i| &object.key[..query.prefix.len() + i + delimiter.len()])
-				} else {
-					None
-				};
-				if let Some(pfx) = common_prefix {
+			// Find if this object has a currently valid (non-deleted,
+			// non-still-uploading) version. If not, skip it.
+			let version = match object.versions().iter().find(|x| x.is_data()) {
+				Some(v) => v,
+				None => continue,
+			};
+
+			// If we don't have space to add this object to our response,
+			// we will need to stop here and mark the key of this object
+			// as the marker from where
+			// we want to start again in the next list call.
+			let cannot_add = result_keys.len() + result_common_prefixes.len() >= query.max_keys;
+
+			// Determine whether this object should be grouped inside
+			// a CommonPrefix because it contains the delimiter,
+			// or if it should be returned as an object.
+			let common_prefix = match &query.delimiter {
+				Some(delimiter) => object.key[query.prefix.len()..]
+					.find(delimiter)
+					.map(|i| &object.key[..query.prefix.len() + i + delimiter.len()]),
+				None => None,
+			};
+			if let Some(pfx) = common_prefix {
+				// In the case where this object must be grouped in a
+				// common prefix, handle it here.
+				if !result_common_prefixes.contains(pfx) {
+					// Determine the first listing key that starts after
+					// the common prefix, by finding the next possible
+					// string by alphabetical order.
+					let mut first_key_after_prefix = pfx.to_string();
+					let tail = first_key_after_prefix.pop().unwrap();
+					first_key_after_prefix.push(((tail as u8) + 1) as char);
+
+					// If this were the end of the chunk,
+					// the next chunk should start after this prefix
+					next_chunk_start = first_key_after_prefix;
+					next_chunk_exclude_start = false;
+
+					if cannot_add {
+						truncated = true;
+						break 'query_loop;
+					}
 					result_common_prefixes.insert(pfx.to_string());
-				} else {
-					let meta = match &version.state {
-						ObjectVersionState::Complete(ObjectVersionData::Inline(meta, _)) => meta,
-						ObjectVersionState::Complete(ObjectVersionData::FirstBlock(meta, _)) => {
-							meta
-						}
-						_ => unreachable!(),
-					};
-					let info = match result_keys.get(&object.key) {
-						None => ListResultInfo {
-							last_modified: version.timestamp,
-							size: meta.size,
-							etag: meta.etag.to_string(),
-						},
-						Some(_lri) => {
-							return Err(Error::InternalError(GarageError::Message(format!(
-								"Duplicate key?? {}",
-								object.key
-							))))
-						}
-					};
-					result_keys.insert(object.key.clone(), info);
-				};
+				}
+				last_processed_item = Some(object.key.clone());
+				continue;
+			};
+
+			// This is not a common prefix, we want to add it to our
+			// response directly.
+			next_chunk_start = object.key.clone();
+
+			if cannot_add {
+				truncated = true;
+				next_chunk_exclude_start = false;
+				break 'query_loop;
 			}
+
+			let meta = match &version.state {
+				ObjectVersionState::Complete(ObjectVersionData::Inline(meta, _)) => meta,
+				ObjectVersionState::Complete(ObjectVersionData::FirstBlock(meta, _)) => meta,
+				_ => unreachable!(),
+			};
+			let info = match result_keys.get(&object.key) {
+				None => ListResultInfo {
+					last_modified: version.timestamp,
+					size: meta.size,
+					etag: meta.etag.to_string(),
+				},
+				Some(_lri) => {
+					return Err(Error::InternalError(GarageError::Message(format!(
+						"Duplicate key?? {} (this is a bug, please report it)",
+						object.key
+					))))
+				}
+			};
+			result_keys.insert(object.key.clone(), info);
+			last_processed_item = Some(object.key.clone());
+			next_chunk_exclude_start = true;
 		}
+
+		// If our database returned less objects than what we were asking for,
+		// it means that no more objects are in the bucket. So we stop here.
 		if objects.len() < query.max_keys + 1 {
-			truncated = None;
+			truncated = false;
 			break 'query_loop;
 		}
-		if !objects.is_empty() {
-			next_chunk_start = objects[objects.len() - 1].key.clone();
+
+		// Sanity check: we should have added at least an object
+		// or a prefix to our returned result.
+		if next_chunk_start == current_chunk_start || last_processed_item.is_none() {
+			return Err(Error::InternalError(GarageError::Message(format!(
+							"S3 ListObject: made no progress, still starting at {} (this is a bug, please report it)", next_chunk_start))));
 		}
+
+		// Loop and fetch more objects
 	}
 
 	let mut result = s3_xml::ListBucketResult {
@@ -181,11 +274,10 @@ pub async fn handle_list(
 			true => Some(s3_xml::Value("url".to_string())),
 			false => None,
 		},
-
 		key_count: Some(s3_xml::IntValue(
 			result_keys.len() as i64 + result_common_prefixes.len() as i64,
 		)),
-		is_truncated: s3_xml::Value(format!("{}", truncated.is_some())),
+		is_truncated: s3_xml::Value(format!("{}", truncated)),
 		contents: vec![],
 		common_prefixes: vec![],
 	};
@@ -197,16 +289,27 @@ pub async fn handle_list(
 		if let Some(sa) = &query.start_after {
 			result.start_after = Some(uriencode_maybe(sa, query.urlencode_resp));
 		}
-		if let Some(nct) = truncated {
-			result.next_continuation_token = Some(s3_xml::Value(base64::encode(nct.as_bytes())));
+		if truncated {
+			let b64 = base64::encode(next_chunk_start.as_bytes());
+			let nct = if next_chunk_exclude_start {
+				format!("]{}", b64)
+			} else {
+				format!("[{}", b64)
+			};
+			result.next_continuation_token = Some(s3_xml::Value(nct));
 		}
 	} else {
 		// TODO: are these supposed to be urlencoded when encoding-type is URL??
 		if let Some(mkr) = &query.marker {
 			result.marker = Some(uriencode_maybe(mkr, query.urlencode_resp));
 		}
-		if let Some(next_marker) = truncated {
-			result.next_marker = Some(uriencode_maybe(&next_marker, query.urlencode_resp));
+		if truncated {
+			if let Some(lpi) = last_processed_item {
+				result.next_marker = Some(uriencode_maybe(&lpi, query.urlencode_resp));
+			} else {
+				return Err(Error::InternalError(GarageError::Message(
+								"S3 ListObject: last_processed_item is None but the response was truncated, indicating that many items were processed (this is a bug, please report it)".to_string())));
+			}
 		}
 	}
 
@@ -221,7 +324,6 @@ pub async fn handle_list(
 	}
 
 	for pfx in result_common_prefixes.iter() {
-		//TODO: in V1, are these urlencoded when urlencode_resp is true ?? (proably)
 		result.common_prefixes.push(s3_xml::CommonPrefix {
 			prefix: uriencode_maybe(pfx, query.urlencode_resp),
 		});

src/api/s3_put.rs

@@ -193,8 +193,8 @@ async fn read_and_put_blocks(
 
 	let mut next_offset = first_block.len();
 	let mut put_curr_version_block = put_block_meta(
-		&garage,
-		&version,
+		garage,
+		version,
 		part_number,
 		0,
 		first_block_hash,
@@ -213,8 +213,8 @@ async fn read_and_put_blocks(
 			let block_hash = blake2sum(&block[..]);
 			let block_len = block.len();
 			put_curr_version_block = put_block_meta(
-				&garage,
-				&version,
+				garage,
+				version,
 				part_number,
 				next_offset as u64,
 				block_hash,
@@ -437,7 +437,7 @@ pub async fn handle_complete_multipart_upload(
 	let body = hyper::body::to_bytes(req.into_body()).await?;
 	verify_signed_content(content_sha256, &body[..])?;
 
-	let body_xml = roxmltree::Document::parse(&std::str::from_utf8(&body)?)?;
+	let body_xml = roxmltree::Document::parse(std::str::from_utf8(&body)?)?;
 	let body_list_of_parts = parse_complete_multpart_upload_body(&body_xml)
 		.ok_or_bad_request("Invalid CompleteMultipartUpload XML")?;
 	debug!(

src/api/signature.rs

@@ -70,7 +70,7 @@ pub async fn check_signature(
 	let canonical_request = canonical_request(
 		request.method(),
 		&request.uri().path().to_string(),
-		&canonical_query_string(&request.uri()),
+		&canonical_query_string(request.uri()),
 		&headers,
 		&authorization.signed_headers,
 		&authorization.content_sha256,
@@ -252,7 +252,7 @@ fn canonical_request(
 		method.as_str(),
 		url_path,
 		canonical_query_string,
-		&canonical_header_string(&headers, signed_headers),
+		&canonical_header_string(headers, signed_headers),
 		"",
 		signed_headers,
 		content_sha256,

src/garage/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "garage"
-version = "0.3.0"
+version = "0.4.0"
 authors = ["Alex Auvolat <alex@adnab.me>"]
 edition = "2018"
 license = "AGPL-3.0"
@@ -14,12 +14,12 @@ path = "main.rs"
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [dependencies]
-garage_api = { version = "0.3.0", path = "../api" }
-garage_model = { version = "0.3.0", path = "../model" }
-garage_rpc = { version = "0.3.0", path = "../rpc" }
-garage_table = { version = "0.3.0", path = "../table" }
-garage_util = { version = "0.3.0", path = "../util" }
-garage_web = { version = "0.3.0", path = "../web" }
+garage_api = { version = "0.4.0", path = "../api" }
+garage_model = { version = "0.4.0", path = "../model" }
+garage_rpc = { version = "0.4.0", path = "../rpc" }
+garage_table = { version = "0.4.0", path = "../table" }
+garage_util = { version = "0.4.0", path = "../util" }
+garage_web = { version = "0.4.0", path = "../web" }
 
 bytes = "1.0"
 git-version = "0.3.4"
@@ -27,6 +27,8 @@ hex = "0.4"
 log = "0.4"
 pretty_env_logger = "0.4"
 rand = "0.8"
+async-trait = "0.1.7"
+sodiumoxide = { version = "0.2.5-0", package = "kuska-sodiumoxide" }
 
 sled = "0.34"
 
@@ -38,3 +40,6 @@ toml = "0.5"
 futures = "0.3"
 futures-util = "0.3"
 tokio = { version = "1.0", default-features = false, features = ["rt", "rt-multi-thread", "io-util", "net", "time", "macros", "sync", "signal", "fs"] }
+
+#netapp = { version = "0.3.0", git = "https://git.deuxfleurs.fr/lx/netapp" }
+netapp = "0.3.0"

src/garage/admin.rs

@@ -2,6 +2,7 @@ use std::collections::HashMap;
 use std::fmt::Write;
 use std::sync::Arc;
 
+use async_trait::async_trait;
 use serde::{Deserialize, Serialize};
 
 use garage_util::error::Error;
@@ -10,8 +11,7 @@ use garage_table::crdt::Crdt;
 use garage_table::replication::*;
 use garage_table::*;
 
-use garage_rpc::rpc_client::*;
-use garage_rpc::rpc_server::*;
+use garage_rpc::*;
 
 use garage_model::bucket_table::*;
 use garage_model::garage::Garage;
@@ -19,10 +19,8 @@ use garage_model::key_table::*;
 
 use crate::cli::*;
 use crate::repair::Repair;
-use crate::*;
 
-pub const ADMIN_RPC_TIMEOUT: Duration = Duration::from_secs(30);
-pub const ADMIN_RPC_PATH: &str = "_admin";
+pub const ADMIN_RPC_PATH: &str = "garage/admin_rpc.rs/Rpc";
 
 #[derive(Debug, Serialize, Deserialize)]
 pub enum AdminRpc {
@@ -39,35 +37,24 @@ pub enum AdminRpc {
 	KeyInfo(Key),
 }
 
-impl RpcMessage for AdminRpc {}
+impl Rpc for AdminRpc {
+	type Response = Result<AdminRpc, Error>;
+}
 
 pub struct AdminRpcHandler {
 	garage: Arc<Garage>,
-	rpc_client: Arc<RpcClient<AdminRpc>>,
+	endpoint: Arc<Endpoint<AdminRpc, Self>>,
 }
 
 impl AdminRpcHandler {
 	pub fn new(garage: Arc<Garage>) -> Arc<Self> {
-		let rpc_client = garage.system.clone().rpc_client::<AdminRpc>(ADMIN_RPC_PATH);
-		Arc::new(Self { garage, rpc_client })
+		let endpoint = garage.system.netapp.endpoint(ADMIN_RPC_PATH.into());
+		let admin = Arc::new(Self { garage, endpoint });
+		admin.endpoint.set_handler(admin.clone());
+		admin
 	}
 
-	pub fn register_handler(self: Arc<Self>, rpc_server: &mut RpcServer) {
-		rpc_server.add_handler::<AdminRpc, _, _>(ADMIN_RPC_PATH.to_string(), move |msg, _addr| {
-			let self2 = self.clone();
-			async move {
-				match msg {
-					AdminRpc::BucketOperation(bo) => self2.handle_bucket_cmd(bo).await,
-					AdminRpc::KeyOperation(ko) => self2.handle_key_cmd(ko).await,
-					AdminRpc::LaunchRepair(opt) => self2.handle_launch_repair(opt).await,
-					AdminRpc::Stats(opt) => self2.handle_stats(opt).await,
-					_ => Err(Error::BadRpc("Invalid RPC".to_string())),
-				}
-			}
-		});
-	}
-
-	async fn handle_bucket_cmd(&self, cmd: BucketOperation) -> Result<AdminRpc, Error> {
+	async fn handle_bucket_cmd(&self, cmd: &BucketOperation) -> Result<AdminRpc, Error> {
 		match cmd {
 			BucketOperation::List => {
 				let bucket_names = self
@@ -187,7 +174,7 @@ impl AdminRpcHandler {
 		}
 	}
 
-	async fn handle_key_cmd(&self, cmd: KeyOperation) -> Result<AdminRpc, Error> {
+	async fn handle_key_cmd(&self, cmd: &KeyOperation) -> Result<AdminRpc, Error> {
 		match cmd {
 			KeyOperation::List => {
 				let key_ids = self
@@ -210,13 +197,13 @@ impl AdminRpcHandler {
 				Ok(AdminRpc::KeyInfo(key))
 			}
 			KeyOperation::New(query) => {
-				let key = Key::new(query.name);
+				let key = Key::new(query.name.clone());
 				self.garage.key_table.insert(&key).await?;
 				Ok(AdminRpc::KeyInfo(key))
 			}
 			KeyOperation::Rename(query) => {
 				let mut key = self.get_existing_key(&query.key_pattern).await?;
-				key.name.update(query.new_name);
+				key.name.update(query.new_name.clone());
 				self.garage.key_table.insert(&key).await?;
 				Ok(AdminRpc::KeyInfo(key))
 			}
@@ -353,17 +340,17 @@ impl AdminRpcHandler {
 			let mut failures = vec![];
 			let ring = self.garage.system.ring.borrow().clone();
 			for node in ring.config.members.keys() {
-				if self
-					.rpc_client
+				let node = (*node).into();
+				let resp = self
+					.endpoint
 					.call(
-						*node,
-						AdminRpc::LaunchRepair(opt_to_send.clone()),
-						ADMIN_RPC_TIMEOUT,
+						&node,
+						&AdminRpc::LaunchRepair(opt_to_send.clone()),
+						PRIO_NORMAL,
 					)
-					.await
-					.is_err()
-				{
-					failures.push(*node);
+					.await;
+				if !matches!(resp, Ok(Ok(_))) {
+					failures.push(node);
 				}
 			}
 			if failures.is_empty() {
@@ -402,10 +389,12 @@ impl AdminRpcHandler {
 
 				writeln!(&mut ret, "\n======================").unwrap();
 				writeln!(&mut ret, "Stats for node {:?}:", node).unwrap();
+
+				let node_id = (*node).into();
 				match self
-					.rpc_client
-					.call(*node, AdminRpc::Stats(opt), ADMIN_RPC_TIMEOUT)
-					.await
+					.endpoint
+					.call(&node_id, &AdminRpc::Stats(opt), PRIO_NORMAL)
+					.await?
 				{
 					Ok(AdminRpc::Ok(s)) => writeln!(&mut ret, "{}", s).unwrap(),
 					Ok(x) => writeln!(&mut ret, "Bad answer: {:?}", x).unwrap(),
@@ -423,7 +412,11 @@ impl AdminRpcHandler {
 		writeln!(
 			&mut ret,
 			"\nGarage version: {}",
-			git_version::git_version!()
+			option_env!("GIT_VERSION").unwrap_or(git_version::git_version!(
+				prefix = "git:",
+				cargo_prefix = "cargo:",
+				fallback = "unknown"
+			))
 		)
 		.unwrap();
 
@@ -492,3 +485,20 @@ impl AdminRpcHandler {
 		writeln!(to, "  GC todo queue length: {}", t.data.gc_todo_len()).unwrap();
 	}
 }
+
+#[async_trait]
+impl EndpointHandler<AdminRpc> for AdminRpcHandler {
+	async fn handle(
+		self: &Arc<Self>,
+		message: &AdminRpc,
+		_from: NodeID,
+	) -> Result<AdminRpc, Error> {
+		match message {
+			AdminRpc::BucketOperation(bo) => self.handle_bucket_cmd(bo).await,
+			AdminRpc::KeyOperation(ko) => self.handle_key_cmd(ko).await,
+			AdminRpc::LaunchRepair(opt) => self.handle_launch_repair(opt.clone()).await,
+			AdminRpc::Stats(opt) => self.handle_stats(opt.clone()).await,
+			_ => Err(Error::BadRpc("Invalid RPC".to_string())),
+		}
+	}
+}

src/garage/cli.rs

@@ -1,597 +0,0 @@
-use std::collections::HashSet;
-use std::net::SocketAddr;
-use std::path::PathBuf;
-
-use serde::{Deserialize, Serialize};
-use structopt::StructOpt;
-
-use garage_util::data::Uuid;
-use garage_util::error::Error;
-use garage_util::time::*;
-
-use garage_rpc::membership::*;
-use garage_rpc::ring::*;
-use garage_rpc::rpc_client::*;
-
-use garage_model::bucket_table::*;
-use garage_model::key_table::*;
-
-use crate::admin_rpc::*;
-
-#[derive(StructOpt, Debug)]
-pub enum Command {
-	/// Run Garage server
-	#[structopt(name = "server")]
-	Server(ServerOpt),
-
-	/// Get network status
-	#[structopt(name = "status")]
-	Status,
-
-	/// Garage node operations
-	#[structopt(name = "node")]
-	Node(NodeOperation),
-
-	/// Bucket operations
-	#[structopt(name = "bucket")]
-	Bucket(BucketOperation),
-
-	/// Key operations
-	#[structopt(name = "key")]
-	Key(KeyOperation),
-
-	/// Start repair of node data
-	#[structopt(name = "repair")]
-	Repair(RepairOpt),
-
-	/// Gather node statistics
-	#[structopt(name = "stats")]
-	Stats(StatsOpt),
-}
-
-#[derive(StructOpt, Debug)]
-pub struct ServerOpt {
-	/// Configuration file
-	#[structopt(short = "c", long = "config", default_value = "./config.toml")]
-	pub config_file: PathBuf,
-}
-
-#[derive(StructOpt, Debug)]
-pub enum NodeOperation {
-	/// Configure Garage node
-	#[structopt(name = "configure")]
-	Configure(ConfigureNodeOpt),
-
-	/// Remove Garage node from cluster
-	#[structopt(name = "remove")]
-	Remove(RemoveNodeOpt),
-}
-
-#[derive(StructOpt, Debug)]
-pub struct ConfigureNodeOpt {
-	/// Node to configure (prefix of hexadecimal node id)
-	node_id: String,
-
-	/// Location (zone or datacenter) of the node
-	#[structopt(short = "z", long = "zone")]
-	zone: Option<String>,
-
-	/// Capacity (in relative terms, use 1 to represent your smallest server)
-	#[structopt(short = "c", long = "capacity")]
-	capacity: Option<u32>,
-
-	/// Gateway-only node
-	#[structopt(short = "g", long = "gateway")]
-	gateway: bool,
-
-	/// Optional node tag
-	#[structopt(short = "t", long = "tag")]
-	tag: Option<String>,
-
-	/// Replaced node(s): list of node IDs that will be removed from the current cluster
-	#[structopt(long = "replace")]
-	replace: Vec<String>,
-}
-
-#[derive(StructOpt, Debug)]
-pub struct RemoveNodeOpt {
-	/// Node to configure (prefix of hexadecimal node id)
-	node_id: String,
-
-	/// If this flag is not given, the node won't be removed
-	#[structopt(long = "yes")]
-	yes: bool,
-}
-
-#[derive(Serialize, Deserialize, StructOpt, Debug)]
-pub enum BucketOperation {
-	/// List buckets
-	#[structopt(name = "list")]
-	List,
-
-	/// Get bucket info
-	#[structopt(name = "info")]
-	Info(BucketOpt),
-
-	/// Create bucket
-	#[structopt(name = "create")]
-	Create(BucketOpt),
-
-	/// Delete bucket
-	#[structopt(name = "delete")]
-	Delete(DeleteBucketOpt),
-
-	/// Allow key to read or write to bucket
-	#[structopt(name = "allow")]
-	Allow(PermBucketOpt),
-
-	/// Allow key to read or write to bucket
-	#[structopt(name = "deny")]
-	Deny(PermBucketOpt),
-
-	/// Expose as website or not
-	#[structopt(name = "website")]
-	Website(WebsiteOpt),
-}
-
-#[derive(Serialize, Deserialize, StructOpt, Debug)]
-pub struct WebsiteOpt {
-	/// Create
-	#[structopt(long = "allow")]
-	pub allow: bool,
-
-	/// Delete
-	#[structopt(long = "deny")]
-	pub deny: bool,
-
-	/// Bucket name
-	pub bucket: String,
-}
-
-#[derive(Serialize, Deserialize, StructOpt, Debug)]
-pub struct BucketOpt {
-	/// Bucket name
-	pub name: String,
-}
-
-#[derive(Serialize, Deserialize, StructOpt, Debug)]
-pub struct DeleteBucketOpt {
-	/// Bucket name
-	pub name: String,
-
-	/// If this flag is not given, the bucket won't be deleted
-	#[structopt(long = "yes")]
-	pub yes: bool,
-}
-
-#[derive(Serialize, Deserialize, StructOpt, Debug)]
-pub struct PermBucketOpt {
-	/// Access key name or ID
-	#[structopt(long = "key")]
-	pub key_pattern: String,
-
-	/// Allow/deny read operations
-	#[structopt(long = "read")]
-	pub read: bool,
-
-	/// Allow/deny write operations
-	#[structopt(long = "write")]
-	pub write: bool,
-
-	/// Bucket name
-	pub bucket: String,
-}
-
-#[derive(Serialize, Deserialize, StructOpt, Debug)]
-pub enum KeyOperation {
-	/// List keys
-	#[structopt(name = "list")]
-	List,
-
-	/// Get key info
-	#[structopt(name = "info")]
-	Info(KeyOpt),
-
-	/// Create new key
-	#[structopt(name = "new")]
-	New(KeyNewOpt),
-
-	/// Rename key
-	#[structopt(name = "rename")]
-	Rename(KeyRenameOpt),
-
-	/// Delete key
-	#[structopt(name = "delete")]
-	Delete(KeyDeleteOpt),
-
-	/// Import key
-	#[structopt(name = "import")]
-	Import(KeyImportOpt),
-}
-
-#[derive(Serialize, Deserialize, StructOpt, Debug)]
-pub struct KeyOpt {
-	/// ID or name of the key
-	pub key_pattern: String,
-}
-
-#[derive(Serialize, Deserialize, StructOpt, Debug)]
-pub struct KeyNewOpt {
-	/// Name of the key
-	#[structopt(long = "name", default_value = "Unnamed key")]
-	pub name: String,
-}
-
-#[derive(Serialize, Deserialize, StructOpt, Debug)]
-pub struct KeyRenameOpt {
-	/// ID or name of the key
-	pub key_pattern: String,
-
-	/// New name of the key
-	pub new_name: String,
-}
-
-#[derive(Serialize, Deserialize, StructOpt, Debug)]
-pub struct KeyDeleteOpt {
-	/// ID or name of the key
-	pub key_pattern: String,
-
-	/// Confirm deletion
-	#[structopt(long = "yes")]
-	pub yes: bool,
-}
-
-#[derive(Serialize, Deserialize, StructOpt, Debug)]
-pub struct KeyImportOpt {
-	/// Access key ID
-	pub key_id: String,
-
-	/// Secret access key
-	pub secret_key: String,
-
-	/// Key name
-	#[structopt(short = "n", default_value = "Imported key")]
-	pub name: String,
-}
-
-#[derive(Serialize, Deserialize, StructOpt, Debug, Clone)]
-pub struct RepairOpt {
-	/// Launch repair operation on all nodes
-	#[structopt(short = "a", long = "all-nodes")]
-	pub all_nodes: bool,
-
-	/// Confirm the launch of the repair operation
-	#[structopt(long = "yes")]
-	pub yes: bool,
-
-	#[structopt(subcommand)]
-	pub what: Option<RepairWhat>,
-}
-
-#[derive(Serialize, Deserialize, StructOpt, Debug, Eq, PartialEq, Clone)]
-pub enum RepairWhat {
-	/// Only do a full sync of metadata tables
-	#[structopt(name = "tables")]
-	Tables,
-	/// Only repair (resync/rebalance) the set of stored blocks
-	#[structopt(name = "blocks")]
-	Blocks,
-	/// Only redo the propagation of object deletions to the version table (slow)
-	#[structopt(name = "versions")]
-	Versions,
-	/// Only redo the propagation of version deletions to the block ref table (extremely slow)
-	#[structopt(name = "block_refs")]
-	BlockRefs,
-}
-
-#[derive(Serialize, Deserialize, StructOpt, Debug, Clone)]
-pub struct StatsOpt {
-	/// Gather statistics from all nodes
-	#[structopt(short = "a", long = "all-nodes")]
-	pub all_nodes: bool,
-
-	/// Gather detailed statistics (this can be long)
-	#[structopt(short = "d", long = "detailed")]
-	pub detailed: bool,
-}
-
-pub async fn cli_cmd(
-	cmd: Command,
-	membership_rpc_cli: RpcAddrClient<Message>,
-	admin_rpc_cli: RpcAddrClient<AdminRpc>,
-	rpc_host: SocketAddr,
-) -> Result<(), Error> {
-	match cmd {
-		Command::Status => cmd_status(membership_rpc_cli, rpc_host).await,
-		Command::Node(NodeOperation::Configure(configure_opt)) => {
-			cmd_configure(membership_rpc_cli, rpc_host, configure_opt).await
-		}
-		Command::Node(NodeOperation::Remove(remove_opt)) => {
-			cmd_remove(membership_rpc_cli, rpc_host, remove_opt).await
-		}
-		Command::Bucket(bo) => {
-			cmd_admin(admin_rpc_cli, rpc_host, AdminRpc::BucketOperation(bo)).await
-		}
-		Command::Key(ko) => cmd_admin(admin_rpc_cli, rpc_host, AdminRpc::KeyOperation(ko)).await,
-		Command::Repair(ro) => cmd_admin(admin_rpc_cli, rpc_host, AdminRpc::LaunchRepair(ro)).await,
-		Command::Stats(so) => cmd_admin(admin_rpc_cli, rpc_host, AdminRpc::Stats(so)).await,
-		_ => unreachable!(),
-	}
-}
-
-pub async fn cmd_status(
-	rpc_cli: RpcAddrClient<Message>,
-	rpc_host: SocketAddr,
-) -> Result<(), Error> {
-	let status = match rpc_cli
-		.call(&rpc_host, &Message::PullStatus, ADMIN_RPC_TIMEOUT)
-		.await??
-	{
-		Message::AdvertiseNodesUp(nodes) => nodes,
-		resp => return Err(Error::Message(format!("Invalid RPC response: {:?}", resp))),
-	};
-	let config = match rpc_cli
-		.call(&rpc_host, &Message::PullConfig, ADMIN_RPC_TIMEOUT)
-		.await??
-	{
-		Message::AdvertiseConfig(cfg) => cfg,
-		resp => return Err(Error::Message(format!("Invalid RPC response: {:?}", resp))),
-	};
-
-	println!("Healthy nodes:");
-	for adv in status.iter().filter(|x| x.is_up) {
-		if let Some(cfg) = config.members.get(&adv.id) {
-			println!(
-				"{:?}\t{}\t{}\t[{}]\t{}\t{}",
-				adv.id,
-				adv.state_info.hostname,
-				adv.addr,
-				cfg.tag,
-				cfg.zone,
-				cfg.capacity_string()
-			);
-		} else {
-			println!(
-				"{:?}\t{}\t{}\tUNCONFIGURED/REMOVED",
-				adv.id, adv.state_info.hostname, adv.addr
-			);
-		}
-	}
-
-	let status_keys = status.iter().map(|x| x.id).collect::<HashSet<_>>();
-	let failure_case_1 = status.iter().any(|x| !x.is_up);
-	let failure_case_2 = config
-		.members
-		.iter()
-		.any(|(id, _)| !status_keys.contains(id));
-	if failure_case_1 || failure_case_2 {
-		println!("\nFailed nodes:");
-		for adv in status.iter().filter(|x| !x.is_up) {
-			if let Some(cfg) = config.members.get(&adv.id) {
-				println!(
-					"{:?}\t{}\t{}\t[{}]\t{}\t{}\tlast seen: {}s ago",
-					adv.id,
-					adv.state_info.hostname,
-					adv.addr,
-					cfg.tag,
-					cfg.zone,
-					cfg.capacity_string(),
-					(now_msec() - adv.last_seen) / 1000,
-				);
-			}
-		}
-		for (id, cfg) in config.members.iter() {
-			if !status.iter().any(|x| x.id == *id) {
-				println!(
-					"{:?}\t{}\t{}\t{}\tnever seen",
-					id,
-					cfg.tag,
-					cfg.zone,
-					cfg.capacity_string(),
-				);
-			}
-		}
-	}
-
-	Ok(())
-}
-
-pub fn find_matching_node(
-	cand: impl std::iter::Iterator<Item = Uuid>,
-	pattern: &str,
-) -> Result<Uuid, Error> {
-	let mut candidates = vec![];
-	for c in cand {
-		if hex::encode(&c).starts_with(&pattern) {
-			candidates.push(c);
-		}
-	}
-	if candidates.len() != 1 {
-		Err(Error::Message(format!(
-			"{} nodes match '{}'",
-			candidates.len(),
-			pattern,
-		)))
-	} else {
-		Ok(candidates[0])
-	}
-}
-
-pub async fn cmd_configure(
-	rpc_cli: RpcAddrClient<Message>,
-	rpc_host: SocketAddr,
-	args: ConfigureNodeOpt,
-) -> Result<(), Error> {
-	let status = match rpc_cli
-		.call(&rpc_host, &Message::PullStatus, ADMIN_RPC_TIMEOUT)
-		.await??
-	{
-		Message::AdvertiseNodesUp(nodes) => nodes,
-		resp => return Err(Error::Message(format!("Invalid RPC response: {:?}", resp))),
-	};
-
-	let added_node = find_matching_node(status.iter().map(|x| x.id), &args.node_id)?;
-
-	let mut config = match rpc_cli
-		.call(&rpc_host, &Message::PullConfig, ADMIN_RPC_TIMEOUT)
-		.await??
-	{
-		Message::AdvertiseConfig(cfg) => cfg,
-		resp => return Err(Error::Message(format!("Invalid RPC response: {:?}", resp))),
-	};
-
-	for replaced in args.replace.iter() {
-		let replaced_node = find_matching_node(config.members.keys().cloned(), replaced)?;
-		if config.members.remove(&replaced_node).is_none() {
-			return Err(Error::Message(format!(
-				"Cannot replace node {:?} as it is not in current configuration",
-				replaced_node
-			)));
-		}
-	}
-
-	if args.capacity.is_some() && args.gateway {
-		return Err(Error::Message(
-				"-c and -g are mutually exclusive, please configure node either with c>0 to act as a storage node or with -g to act as a gateway node".into()));
-	}
-	if args.capacity == Some(0) {
-		return Err(Error::Message("Invalid capacity value: 0".into()));
-	}
-
-	let new_entry = match config.members.get(&added_node) {
-		None => {
-			let capacity = match args.capacity {
-				Some(c) => Some(c),
-				None if args.gateway => None,
-				_ => return Err(Error::Message(
-						"Please specify a capacity with the -c flag, or set node explicitly as gateway with -g".into())),
-			};
-			NetworkConfigEntry {
-				zone: args.zone.expect("Please specifiy a zone with the -z flag"),
-				capacity,
-				tag: args.tag.unwrap_or_default(),
-			}
-		}
-		Some(old) => {
-			let capacity = match args.capacity {
-				Some(c) => Some(c),
-				None if args.gateway => None,
-				_ => old.capacity,
-			};
-			NetworkConfigEntry {
-				zone: args.zone.unwrap_or_else(|| old.zone.to_string()),
-				capacity,
-				tag: args.tag.unwrap_or_else(|| old.tag.to_string()),
-			}
-		}
-	};
-
-	config.members.insert(added_node, new_entry);
-	config.version += 1;
-
-	rpc_cli
-		.call(
-			&rpc_host,
-			&Message::AdvertiseConfig(config),
-			ADMIN_RPC_TIMEOUT,
-		)
-		.await??;
-	Ok(())
-}
-
-pub async fn cmd_remove(
-	rpc_cli: RpcAddrClient<Message>,
-	rpc_host: SocketAddr,
-	args: RemoveNodeOpt,
-) -> Result<(), Error> {
-	let mut config = match rpc_cli
-		.call(&rpc_host, &Message::PullConfig, ADMIN_RPC_TIMEOUT)
-		.await??
-	{
-		Message::AdvertiseConfig(cfg) => cfg,
-		resp => return Err(Error::Message(format!("Invalid RPC response: {:?}", resp))),
-	};
-
-	let deleted_node = find_matching_node(config.members.keys().cloned(), &args.node_id)?;
-
-	if !args.yes {
-		return Err(Error::Message(format!(
-			"Add the flag --yes to really remove {:?} from the cluster",
-			deleted_node
-		)));
-	}
-
-	config.members.remove(&deleted_node);
-	config.version += 1;
-
-	rpc_cli
-		.call(
-			&rpc_host,
-			&Message::AdvertiseConfig(config),
-			ADMIN_RPC_TIMEOUT,
-		)
-		.await??;
-	Ok(())
-}
-
-pub async fn cmd_admin(
-	rpc_cli: RpcAddrClient<AdminRpc>,
-	rpc_host: SocketAddr,
-	args: AdminRpc,
-) -> Result<(), Error> {
-	match rpc_cli.call(&rpc_host, args, ADMIN_RPC_TIMEOUT).await?? {
-		AdminRpc::Ok(msg) => {
-			println!("{}", msg);
-		}
-		AdminRpc::BucketList(bl) => {
-			println!("List of buckets:");
-			for bucket in bl {
-				println!("{}", bucket);
-			}
-		}
-		AdminRpc::BucketInfo(bucket) => {
-			print_bucket_info(&bucket);
-		}
-		AdminRpc::KeyList(kl) => {
-			println!("List of keys:");
-			for key in kl {
-				println!("{}\t{}", key.0, key.1);
-			}
-		}
-		AdminRpc::KeyInfo(key) => {
-			print_key_info(&key);
-		}
-		r => {
-			error!("Unexpected response: {:?}", r);
-		}
-	}
-	Ok(())
-}
-
-fn print_key_info(key: &Key) {
-	println!("Key name: {}", key.name.get());
-	println!("Key ID: {}", key.key_id);
-	println!("Secret key: {}", key.secret_key);
-	if key.deleted.get() {
-		println!("Key is deleted.");
-	} else {
-		println!("Authorized buckets:");
-		for (b, _, perm) in key.authorized_buckets.items().iter() {
-			println!("- {} R:{} W:{}", b, perm.allow_read, perm.allow_write);
-		}
-	}
-}
-
-fn print_bucket_info(bucket: &Bucket) {
-	println!("Bucket name: {}", bucket.name);
-	match bucket.state.get() {
-		BucketState::Deleted => println!("Bucket is deleted."),
-		BucketState::Present(p) => {
-			println!("Authorized keys:");
-			for (k, _, perm) in p.authorized_keys.items().iter() {
-				println!("- {} R:{} W:{}", k, perm.allow_read, perm.allow_write);
-			}
-			println!("Website access: {}", p.website.get());
-		}
-	};
-}

src/garage/cli/cmd.rs

@@ -0,0 +1,287 @@
+use std::collections::HashSet;
+
+use garage_util::error::*;
+
+use garage_rpc::ring::*;
+use garage_rpc::system::*;
+use garage_rpc::*;
+
+use crate::admin::*;
+use crate::cli::*;
+
+pub async fn cli_command_dispatch(
+	cmd: Command,
+	system_rpc_endpoint: &Endpoint<SystemRpc, ()>,
+	admin_rpc_endpoint: &Endpoint<AdminRpc, ()>,
+	rpc_host: NodeID,
+) -> Result<(), Error> {
+	match cmd {
+		Command::Status => cmd_status(system_rpc_endpoint, rpc_host).await,
+		Command::Node(NodeOperation::Connect(connect_opt)) => {
+			cmd_connect(system_rpc_endpoint, rpc_host, connect_opt).await
+		}
+		Command::Node(NodeOperation::Configure(configure_opt)) => {
+			cmd_configure(system_rpc_endpoint, rpc_host, configure_opt).await
+		}
+		Command::Node(NodeOperation::Remove(remove_opt)) => {
+			cmd_remove(system_rpc_endpoint, rpc_host, remove_opt).await
+		}
+		Command::Bucket(bo) => {
+			cmd_admin(admin_rpc_endpoint, rpc_host, AdminRpc::BucketOperation(bo)).await
+		}
+		Command::Key(ko) => {
+			cmd_admin(admin_rpc_endpoint, rpc_host, AdminRpc::KeyOperation(ko)).await
+		}
+		Command::Repair(ro) => {
+			cmd_admin(admin_rpc_endpoint, rpc_host, AdminRpc::LaunchRepair(ro)).await
+		}
+		Command::Stats(so) => cmd_admin(admin_rpc_endpoint, rpc_host, AdminRpc::Stats(so)).await,
+		_ => unreachable!(),
+	}
+}
+
+pub async fn cmd_status(rpc_cli: &Endpoint<SystemRpc, ()>, rpc_host: NodeID) -> Result<(), Error> {
+	let status = match rpc_cli
+		.call(&rpc_host, &SystemRpc::GetKnownNodes, PRIO_NORMAL)
+		.await??
+	{
+		SystemRpc::ReturnKnownNodes(nodes) => nodes,
+		resp => return Err(Error::Message(format!("Invalid RPC response: {:?}", resp))),
+	};
+	let config = match rpc_cli
+		.call(&rpc_host, &SystemRpc::PullConfig, PRIO_NORMAL)
+		.await??
+	{
+		SystemRpc::AdvertiseConfig(cfg) => cfg,
+		resp => return Err(Error::Message(format!("Invalid RPC response: {:?}", resp))),
+	};
+
+	println!("==== HEALTHY NODES ====");
+	let mut healthy_nodes = vec!["ID\tHostname\tAddress\tTag\tZone\tCapacity".to_string()];
+	for adv in status.iter().filter(|adv| adv.is_up) {
+		if let Some(cfg) = config.members.get(&adv.id) {
+			healthy_nodes.push(format!(
+				"{id:?}\t{host}\t{addr}\t[{tag}]\t{zone}\t{capacity}",
+				id = adv.id,
+				host = adv.status.hostname,
+				addr = adv.addr,
+				tag = cfg.tag,
+				zone = cfg.zone,
+				capacity = cfg.capacity_string(),
+			));
+		} else {
+			healthy_nodes.push(format!(
+				"{id:?}\t{h}\t{addr}\tNO ROLE ASSIGNED",
+				id = adv.id,
+				h = adv.status.hostname,
+				addr = adv.addr,
+			));
+		}
+	}
+	format_table(healthy_nodes);
+
+	let status_keys = status.iter().map(|adv| adv.id).collect::<HashSet<_>>();
+	let failure_case_1 = status.iter().any(|adv| !adv.is_up);
+	let failure_case_2 = config
+		.members
+		.iter()
+		.any(|(id, _)| !status_keys.contains(id));
+	if failure_case_1 || failure_case_2 {
+		println!("\n==== FAILED NODES ====");
+		let mut failed_nodes =
+			vec!["ID\tHostname\tAddress\tTag\tZone\tCapacity\tLast seen".to_string()];
+		for adv in status.iter().filter(|adv| !adv.is_up) {
+			if let Some(cfg) = config.members.get(&adv.id) {
+				failed_nodes.push(format!(
+					"{id:?}\t{host}\t{addr}\t[{tag}]\t{zone}\t{capacity}\t{last_seen}",
+					id = adv.id,
+					host = adv.status.hostname,
+					addr = adv.addr,
+					tag = cfg.tag,
+					zone = cfg.zone,
+					capacity = cfg.capacity_string(),
+					last_seen = adv
+						.last_seen_secs_ago
+						.map(|s| format!("{}s ago", s))
+						.unwrap_or_else(|| "never seen".into()),
+				));
+			}
+		}
+		for (id, cfg) in config.members.iter() {
+			if !status_keys.contains(id) {
+				failed_nodes.push(format!(
+					"{id:?}\t??\t??\t[{tag}]\t{zone}\t{capacity}\tnever seen",
+					id = id,
+					tag = cfg.tag,
+					zone = cfg.zone,
+					capacity = cfg.capacity_string(),
+				));
+			}
+		}
+		format_table(failed_nodes);
+	}
+
+	Ok(())
+}
+
+pub async fn cmd_connect(
+	rpc_cli: &Endpoint<SystemRpc, ()>,
+	rpc_host: NodeID,
+	args: ConnectNodeOpt,
+) -> Result<(), Error> {
+	match rpc_cli
+		.call(&rpc_host, &SystemRpc::Connect(args.node), PRIO_NORMAL)
+		.await??
+	{
+		SystemRpc::Ok => {
+			println!("Success.");
+			Ok(())
+		}
+		r => Err(Error::BadRpc(format!("Unexpected response: {:?}", r))),
+	}
+}
+
+pub async fn cmd_configure(
+	rpc_cli: &Endpoint<SystemRpc, ()>,
+	rpc_host: NodeID,
+	args: ConfigureNodeOpt,
+) -> Result<(), Error> {
+	let status = match rpc_cli
+		.call(&rpc_host, &SystemRpc::GetKnownNodes, PRIO_NORMAL)
+		.await??
+	{
+		SystemRpc::ReturnKnownNodes(nodes) => nodes,
+		resp => return Err(Error::Message(format!("Invalid RPC response: {:?}", resp))),
+	};
+
+	let added_node = find_matching_node(status.iter().map(|adv| adv.id), &args.node_id)?;
+
+	let mut config = match rpc_cli
+		.call(&rpc_host, &SystemRpc::PullConfig, PRIO_NORMAL)
+		.await??
+	{
+		SystemRpc::AdvertiseConfig(cfg) => cfg,
+		resp => return Err(Error::Message(format!("Invalid RPC response: {:?}", resp))),
+	};
+
+	for replaced in args.replace.iter() {
+		let replaced_node = find_matching_node(config.members.keys().cloned(), replaced)?;
+		if config.members.remove(&replaced_node).is_none() {
+			return Err(Error::Message(format!(
+				"Cannot replace node {:?} as it is not in current configuration",
+				replaced_node
+			)));
+		}
+	}
+
+	if args.capacity.is_some() && args.gateway {
+		return Err(Error::Message(
+				"-c and -g are mutually exclusive, please configure node either with c>0 to act as a storage node or with -g to act as a gateway node".into()));
+	}
+	if args.capacity == Some(0) {
+		return Err(Error::Message("Invalid capacity value: 0".into()));
+	}
+
+	let new_entry = match config.members.get(&added_node) {
+		None => {
+			let capacity = match args.capacity {
+				Some(c) => Some(c),
+				None if args.gateway => None,
+				_ => return Err(Error::Message(
+						"Please specify a capacity with the -c flag, or set node explicitly as gateway with -g".into())),
+			};
+			NetworkConfigEntry {
+				zone: args.zone.ok_or("Please specifiy a zone with the -z flag")?,
+				capacity,
+				tag: args.tag.unwrap_or_default(),
+			}
+		}
+		Some(old) => {
+			let capacity = match args.capacity {
+				Some(c) => Some(c),
+				None if args.gateway => None,
+				_ => old.capacity,
+			};
+			NetworkConfigEntry {
+				zone: args.zone.unwrap_or_else(|| old.zone.to_string()),
+				capacity,
+				tag: args.tag.unwrap_or_else(|| old.tag.to_string()),
+			}
+		}
+	};
+
+	config.members.insert(added_node, new_entry);
+	config.version += 1;
+
+	rpc_cli
+		.call(&rpc_host, &SystemRpc::AdvertiseConfig(config), PRIO_NORMAL)
+		.await??;
+	Ok(())
+}
+
+pub async fn cmd_remove(
+	rpc_cli: &Endpoint<SystemRpc, ()>,
+	rpc_host: NodeID,
+	args: RemoveNodeOpt,
+) -> Result<(), Error> {
+	let mut config = match rpc_cli
+		.call(&rpc_host, &SystemRpc::PullConfig, PRIO_NORMAL)
+		.await??
+	{
+		SystemRpc::AdvertiseConfig(cfg) => cfg,
+		resp => return Err(Error::Message(format!("Invalid RPC response: {:?}", resp))),
+	};
+
+	let deleted_node = find_matching_node(config.members.keys().cloned(), &args.node_id)?;
+
+	if !args.yes {
+		return Err(Error::Message(format!(
+			"Add the flag --yes to really remove {:?} from the cluster",
+			deleted_node
+		)));
+	}
+
+	config.members.remove(&deleted_node);
+	config.version += 1;
+
+	rpc_cli
+		.call(&rpc_host, &SystemRpc::AdvertiseConfig(config), PRIO_NORMAL)
+		.await??;
+	Ok(())
+}
+
+pub async fn cmd_admin(
+	rpc_cli: &Endpoint<AdminRpc, ()>,
+	rpc_host: NodeID,
+	args: AdminRpc,
+) -> Result<(), Error> {
+	match rpc_cli.call(&rpc_host, &args, PRIO_NORMAL).await?? {
+		AdminRpc::Ok(msg) => {
+			println!("{}", msg);
+		}
+		AdminRpc::BucketList(bl) => {
+			println!("List of buckets:");
+			for bucket in bl {
+				println!("{}", bucket);
+			}
+		}
+		AdminRpc::BucketInfo(bucket) => {
+			print_bucket_info(&bucket);
+		}
+		AdminRpc::KeyList(kl) => {
+			println!("List of keys:");
+			for key in kl {
+				println!("{}\t{}", key.0, key.1);
+			}
+		}
+		AdminRpc::KeyInfo(key) => {
+			print_key_info(&key);
+		}
+		r => {
+			error!("Unexpected response: {:?}", r);
+		}
+	}
+	Ok(())
+}
+
+// --- Utility functions ----

src/garage/cli/init.rs

@@ -0,0 +1,67 @@
+use std::path::PathBuf;
+
+use log::warn;
+
+use garage_util::error::*;
+
+pub const READ_KEY_ERROR: &str = "Unable to read node key. It will be generated by your garage node the first time is it launched. Ensure that your garage node is currently running. (The node key is supposed to be stored in your metadata directory.)";
+
+pub fn node_id_command(config_file: PathBuf, quiet: bool) -> Result<(), Error> {
+	let config = garage_util::config::read_config(config_file.clone()).err_context(format!(
+		"Unable to read configuration file {}",
+		config_file.to_string_lossy(),
+	))?;
+
+	let node_id =
+		garage_rpc::system::read_node_id(&config.metadata_dir).err_context(READ_KEY_ERROR)?;
+
+	let idstr = if let Some(addr) = config.rpc_public_addr {
+		let idstr = format!("{}@{}", hex::encode(&node_id), addr);
+		println!("{}", idstr);
+		idstr
+	} else {
+		let idstr = hex::encode(&node_id);
+		println!("{}", idstr);
+
+		if !quiet {
+			warn!("WARNING: I don't know the public address to reach this node.");
+			warn!("In all of the instructions below, replace 127.0.0.1:{} by the appropriate address and port.", config.rpc_bind_addr.port());
+		}
+
+		format!("{}@127.0.0.1:{}", idstr, config.rpc_bind_addr.port())
+	};
+
+	if !quiet {
+		eprintln!();
+		eprintln!(
+			"To instruct a node to connect to this node, run the following command on that node:"
+		);
+		eprintln!("    garage [-c <config file path>] node connect {}", idstr);
+		eprintln!();
+		eprintln!("Or instruct them to connect from here by running:");
+		eprintln!(
+			"    garage -c {} -h <remote node> node connect {}",
+			config_file.to_string_lossy(),
+			idstr
+		);
+		eprintln!(
+			"where <remote_node> is their own node identifier in the format: <pubkey>@<ip>:<port>"
+		);
+		eprintln!();
+		eprintln!("This node identifier can also be added as a bootstrap node in other node's garage.toml files:");
+		eprintln!("    bootstrap_peers = [");
+		eprintln!("        \"{}\",", idstr);
+		eprintln!("        ...");
+		eprintln!("    ]");
+		eprintln!();
+		eprintln!(
+			r#"Security notice: Garage's intra-cluster communications are secured primarily by the shared
+secret value rpc_secret.  However, an attacker that knows rpc_secret (for example if it
+leaks) cannot connect if they do not know any of the identifiers of the nodes in the
+cluster. It is thus a good security measure to try to keep them secret if possible.
+				  "#
+		);
+	}
+
+	Ok(())
+}

src/garage/cli/mod.rs

@@ -0,0 +1,9 @@
+pub(crate) mod cmd;
+pub(crate) mod init;
+pub(crate) mod structs;
+pub(crate) mod util;
+
+pub(crate) use cmd::*;
+pub(crate) use init::*;
+pub(crate) use structs::*;
+pub(crate) use util::*;

src/garage/cli/structs.rs

@@ -0,0 +1,303 @@
+use serde::{Deserialize, Serialize};
+
+use structopt::StructOpt;
+
+#[derive(StructOpt, Debug)]
+pub enum Command {
+	/// Run Garage server
+	#[structopt(name = "server")]
+	Server,
+
+	/// Print identifier (public key) of this garage node.
+	/// Generates a new keypair if necessary.
+	#[structopt(name = "node-id")]
+	NodeId(NodeIdOpt),
+
+	/// Get network status
+	#[structopt(name = "status")]
+	Status,
+
+	/// Garage node operations
+	#[structopt(name = "node")]
+	Node(NodeOperation),
+
+	/// Bucket operations
+	#[structopt(name = "bucket")]
+	Bucket(BucketOperation),
+
+	/// Key operations
+	#[structopt(name = "key")]
+	Key(KeyOperation),
+
+	/// Start repair of node data
+	#[structopt(name = "repair")]
+	Repair(RepairOpt),
+
+	/// Gather node statistics
+	#[structopt(name = "stats")]
+	Stats(StatsOpt),
+}
+
+#[derive(StructOpt, Debug)]
+pub enum NodeOperation {
+	/// Connect to Garage node that is currently isolated from the system
+	#[structopt(name = "connect")]
+	Connect(ConnectNodeOpt),
+
+	/// Configure Garage node
+	#[structopt(name = "configure")]
+	Configure(ConfigureNodeOpt),
+
+	/// Remove Garage node from cluster
+	#[structopt(name = "remove")]
+	Remove(RemoveNodeOpt),
+}
+
+#[derive(StructOpt, Debug)]
+pub struct NodeIdOpt {
+	/// Do not print usage instructions to stderr
+	#[structopt(short = "q", long = "quiet")]
+	pub(crate) quiet: bool,
+}
+
+#[derive(StructOpt, Debug)]
+pub struct ConnectNodeOpt {
+	/// Node public key and address, in the format:
+	/// `<public key hexadecimal>@<ip or hostname>:<port>`
+	pub(crate) node: String,
+}
+
+#[derive(StructOpt, Debug)]
+pub struct ConfigureNodeOpt {
+	/// Node to configure (prefix of hexadecimal node id)
+	pub(crate) node_id: String,
+
+	/// Location (zone or datacenter) of the node
+	#[structopt(short = "z", long = "zone")]
+	pub(crate) zone: Option<String>,
+
+	/// Capacity (in relative terms, use 1 to represent your smallest server)
+	#[structopt(short = "c", long = "capacity")]
+	pub(crate) capacity: Option<u32>,
+
+	/// Gateway-only node
+	#[structopt(short = "g", long = "gateway")]
+	pub(crate) gateway: bool,
+
+	/// Optional node tag
+	#[structopt(short = "t", long = "tag")]
+	pub(crate) tag: Option<String>,
+
+	/// Replaced node(s): list of node IDs that will be removed from the current cluster
+	#[structopt(long = "replace")]
+	pub(crate) replace: Vec<String>,
+}
+
+#[derive(StructOpt, Debug)]
+pub struct RemoveNodeOpt {
+	/// Node to configure (prefix of hexadecimal node id)
+	pub(crate) node_id: String,
+
+	/// If this flag is not given, the node won't be removed
+	#[structopt(long = "yes")]
+	pub(crate) yes: bool,
+}
+
+#[derive(Serialize, Deserialize, StructOpt, Debug)]
+pub enum BucketOperation {
+	/// List buckets
+	#[structopt(name = "list")]
+	List,
+
+	/// Get bucket info
+	#[structopt(name = "info")]
+	Info(BucketOpt),
+
+	/// Create bucket
+	#[structopt(name = "create")]
+	Create(BucketOpt),
+
+	/// Delete bucket
+	#[structopt(name = "delete")]
+	Delete(DeleteBucketOpt),
+
+	/// Allow key to read or write to bucket
+	#[structopt(name = "allow")]
+	Allow(PermBucketOpt),
+
+	/// Deny key from reading or writing to bucket
+	#[structopt(name = "deny")]
+	Deny(PermBucketOpt),
+
+	/// Expose as website or not
+	#[structopt(name = "website")]
+	Website(WebsiteOpt),
+}
+
+#[derive(Serialize, Deserialize, StructOpt, Debug)]
+pub struct WebsiteOpt {
+	/// Create
+	#[structopt(long = "allow")]
+	pub allow: bool,
+
+	/// Delete
+	#[structopt(long = "deny")]
+	pub deny: bool,
+
+	/// Bucket name
+	pub bucket: String,
+}
+
+#[derive(Serialize, Deserialize, StructOpt, Debug)]
+pub struct BucketOpt {
+	/// Bucket name
+	pub name: String,
+}
+
+#[derive(Serialize, Deserialize, StructOpt, Debug)]
+pub struct DeleteBucketOpt {
+	/// Bucket name
+	pub name: String,
+
+	/// If this flag is not given, the bucket won't be deleted
+	#[structopt(long = "yes")]
+	pub yes: bool,
+}
+
+#[derive(Serialize, Deserialize, StructOpt, Debug)]
+pub struct PermBucketOpt {
+	/// Access key name or ID
+	#[structopt(long = "key")]
+	pub key_pattern: String,
+
+	/// Allow/deny read operations
+	#[structopt(long = "read")]
+	pub read: bool,
+
+	/// Allow/deny write operations
+	#[structopt(long = "write")]
+	pub write: bool,
+
+	/// Bucket name
+	pub bucket: String,
+}
+
+#[derive(Serialize, Deserialize, StructOpt, Debug)]
+pub enum KeyOperation {
+	/// List keys
+	#[structopt(name = "list")]
+	List,
+
+	/// Get key info
+	#[structopt(name = "info")]
+	Info(KeyOpt),
+
+	/// Create new key
+	#[structopt(name = "new")]
+	New(KeyNewOpt),
+
+	/// Rename key
+	#[structopt(name = "rename")]
+	Rename(KeyRenameOpt),
+
+	/// Delete key
+	#[structopt(name = "delete")]
+	Delete(KeyDeleteOpt),
+
+	/// Import key
+	#[structopt(name = "import")]
+	Import(KeyImportOpt),
+}
+
+#[derive(Serialize, Deserialize, StructOpt, Debug)]
+pub struct KeyOpt {
+	/// ID or name of the key
+	pub key_pattern: String,
+}
+
+#[derive(Serialize, Deserialize, StructOpt, Debug)]
+pub struct KeyNewOpt {
+	/// Name of the key
+	#[structopt(long = "name", default_value = "Unnamed key")]
+	pub name: String,
+}
+
+#[derive(Serialize, Deserialize, StructOpt, Debug)]
+pub struct KeyRenameOpt {
+	/// ID or name of the key
+	pub key_pattern: String,
+
+	/// New name of the key
+	pub new_name: String,
+}
+
+#[derive(Serialize, Deserialize, StructOpt, Debug)]
+pub struct KeyDeleteOpt {
+	/// ID or name of the key
+	pub key_pattern: String,
+
+	/// Confirm deletion
+	#[structopt(long = "yes")]
+	pub yes: bool,
+}
+
+#[derive(Serialize, Deserialize, StructOpt, Debug)]
+pub struct KeyImportOpt {
+	/// Access key ID
+	pub key_id: String,
+
+	/// Secret access key
+	pub secret_key: String,
+
+	/// Key name
+	#[structopt(short = "n", default_value = "Imported key")]
+	pub name: String,
+}
+
+#[derive(Serialize, Deserialize, StructOpt, Debug, Clone)]
+pub struct RepairOpt {
+	/// Launch repair operation on all nodes
+	#[structopt(short = "a", long = "all-nodes")]
+	pub all_nodes: bool,
+
+	/// Confirm the launch of the repair operation
+	#[structopt(long = "yes")]
+	pub yes: bool,
+
+	#[structopt(subcommand)]
+	pub what: RepairWhat,
+}
+
+#[derive(Serialize, Deserialize, StructOpt, Debug, Eq, PartialEq, Clone)]
+pub enum RepairWhat {
+	/// Only do a full sync of metadata tables
+	#[structopt(name = "tables")]
+	Tables,
+	/// Only repair (resync/rebalance) the set of stored blocks
+	#[structopt(name = "blocks")]
+	Blocks,
+	/// Only redo the propagation of object deletions to the version table (slow)
+	#[structopt(name = "versions")]
+	Versions,
+	/// Only redo the propagation of version deletions to the block ref table (extremely slow)
+	#[structopt(name = "block_refs")]
+	BlockRefs,
+	/// Verify integrity of all blocks on disc (extremely slow, i/o intensive)
+	#[structopt(name = "scrub")]
+	Scrub {
+		/// Tranquility factor (see tranquilizer documentation)
+		#[structopt(name = "tranquility", default_value = "2")]
+		tranquility: u32,
+	},
+}
+
+#[derive(Serialize, Deserialize, StructOpt, Debug, Clone)]
+pub struct StatsOpt {
+	/// Gather statistics from all nodes
+	#[structopt(short = "a", long = "all-nodes")]
+	pub all_nodes: bool,
+
+	/// Gather detailed statistics (this can be long)
+	#[structopt(short = "d", long = "detailed")]
+	pub detailed: bool,
+}

src/garage/cli/util.rs

@@ -0,0 +1,83 @@
+use garage_util::data::Uuid;
+use garage_util::error::*;
+
+use garage_model::bucket_table::*;
+use garage_model::key_table::*;
+
+pub fn print_key_info(key: &Key) {
+	println!("Key name: {}", key.name.get());
+	println!("Key ID: {}", key.key_id);
+	println!("Secret key: {}", key.secret_key);
+	if key.deleted.get() {
+		println!("Key is deleted.");
+	} else {
+		println!("Authorized buckets:");
+		for (b, _, perm) in key.authorized_buckets.items().iter() {
+			println!("- {} R:{} W:{}", b, perm.allow_read, perm.allow_write);
+		}
+	}
+}
+
+pub fn print_bucket_info(bucket: &Bucket) {
+	println!("Bucket name: {}", bucket.name);
+	match bucket.state.get() {
+		BucketState::Deleted => println!("Bucket is deleted."),
+		BucketState::Present(p) => {
+			println!("Authorized keys:");
+			for (k, _, perm) in p.authorized_keys.items().iter() {
+				println!("- {} R:{} W:{}", k, perm.allow_read, perm.allow_write);
+			}
+			println!("Website access: {}", p.website.get());
+		}
+	};
+}
+
+pub fn format_table(data: Vec<String>) {
+	let data = data
+		.iter()
+		.map(|s| s.split('\t').collect::<Vec<_>>())
+		.collect::<Vec<_>>();
+
+	let columns = data.iter().map(|row| row.len()).fold(0, std::cmp::max);
+	let mut column_size = vec![0; columns];
+
+	let mut out = String::new();
+
+	for row in data.iter() {
+		for (i, col) in row.iter().enumerate() {
+			column_size[i] = std::cmp::max(column_size[i], col.chars().count());
+		}
+	}
+
+	for row in data.iter() {
+		for (col, col_len) in row[..row.len() - 1].iter().zip(column_size.iter()) {
+			out.push_str(col);
+			(0..col_len - col.chars().count() + 2).for_each(|_| out.push(' '));
+		}
+		out.push_str(row[row.len() - 1]);
+		out.push('\n');
+	}
+
+	print!("{}", out);
+}
+
+pub fn find_matching_node(
+	cand: impl std::iter::Iterator<Item = Uuid>,
+	pattern: &str,
+) -> Result<Uuid, Error> {
+	let mut candidates = vec![];
+	for c in cand {
+		if hex::encode(&c).starts_with(&pattern) {
+			candidates.push(c);
+		}
+	}
+	if candidates.len() != 1 {
+		Err(Error::Message(format!(
+			"{} nodes match '{}'",
+			candidates.len(),
+			pattern,
+		)))
+	} else {
+		Ok(candidates[0])
+	}
+}

src/garage/main.rs

@@ -4,39 +4,47 @@
 #[macro_use]
 extern crate log;
 
-mod admin_rpc;
+mod admin;
 mod cli;
 mod repair;
 mod server;
 
 use std::net::SocketAddr;
-use std::sync::Arc;
-use std::time::Duration;
+use std::path::PathBuf;
 
 use structopt::StructOpt;
 
-use garage_util::config::TlsConfig;
-use garage_util::error::Error;
+use netapp::util::parse_and_resolve_peer_addr;
+use netapp::NetworkKey;
 
-use garage_rpc::membership::*;
-use garage_rpc::rpc_client::*;
+use garage_util::error::*;
 
-use admin_rpc::*;
+use garage_rpc::system::*;
+use garage_rpc::*;
+
+use admin::*;
 use cli::*;
 
 #[derive(StructOpt, Debug)]
 #[structopt(name = "garage")]
 struct Opt {
-	/// RPC connect to this host to execute client operations
-	#[structopt(short = "h", long = "rpc-host", default_value = "127.0.0.1:3901")]
-	pub rpc_host: SocketAddr,
+	/// Host to connect to for admin operations, in the format:
+	/// <public-key>@<ip>:<port>
+	#[structopt(short = "h", long = "rpc-host", env = "GARAGE_RPC_HOST")]
+	pub rpc_host: Option<String>,
 
-	#[structopt(long = "ca-cert")]
-	pub ca_cert: Option<String>,
-	#[structopt(long = "client-cert")]
-	pub client_cert: Option<String>,
-	#[structopt(long = "client-key")]
-	pub client_key: Option<String>,
+	/// RPC secret network key for admin operations
+	#[structopt(short = "s", long = "rpc-secret", env = "GARAGE_RPC_SECRET")]
+	pub rpc_secret: Option<String>,
+
+	/// Configuration file (garage.toml)
+	#[structopt(
+		short = "c",
+		long = "config",
+		env = "GARAGE_CONFIG_FILE",
+		default_value = "/etc/garage.toml"
+	)]
+	pub config_file: PathBuf,
 
 	#[structopt(subcommand)]
 	cmd: Command,
@@ -44,46 +52,87 @@ struct Opt {
 
 #[tokio::main]
 async fn main() {
+	if std::env::var("RUST_LOG").is_err() {
+		std::env::set_var("RUST_LOG", "garage=info")
+	}
 	pretty_env_logger::init();
+	sodiumoxide::init().expect("Unable to init sodiumoxide");
 
 	let opt = Opt::from_args();
 
-	let res = if let Command::Server(server_opt) = opt.cmd {
-		// Abort on panic (same behavior as in Go)
-		std::panic::set_hook(Box::new(|panic_info| {
-			error!("{}", panic_info.to_string());
-			std::process::abort();
-		}));
+	let res = match opt.cmd {
+		Command::Server => {
+			// Abort on panic (same behavior as in Go)
+			std::panic::set_hook(Box::new(|panic_info| {
+				error!("{}", panic_info.to_string());
+				std::process::abort();
+			}));
 
-		server::run_server(server_opt.config_file).await
-	} else {
-		cli_command(opt).await
+			server::run_server(opt.config_file).await
+		}
+		Command::NodeId(node_id_opt) => node_id_command(opt.config_file, node_id_opt.quiet),
+		_ => cli_command(opt).await,
 	};
 
 	if let Err(e) = res {
-		error!("{}", e);
+		eprintln!("Error: {}", e);
+		std::process::exit(1);
 	}
 }
 
 async fn cli_command(opt: Opt) -> Result<(), Error> {
-	let tls_config = match (opt.ca_cert, opt.client_cert, opt.client_key) {
-		(Some(ca_cert), Some(client_cert), Some(client_key)) => Some(TlsConfig {
-			ca_cert,
-			node_cert: client_cert,
-			node_key: client_key,
-		}),
-		(None, None, None) => None,
-		_ => {
-			warn!("Missing one of: --ca-cert, --node-cert, --node-key. Not using TLS.");
-			None
+	let config = if opt.rpc_secret.is_none() || opt.rpc_host.is_none() {
+		Some(garage_util::config::read_config(opt.config_file.clone())
+			.err_context(format!("Unable to read configuration file {}. Configuration file is needed because -h or -s is not provided on the command line.", opt.config_file.to_string_lossy()))?)
+	} else {
+		None
+	};
+
+	// Find and parse network RPC secret
+	let net_key_hex_str = opt
+		.rpc_secret
+		.as_ref()
+		.or_else(|| config.as_ref().map(|c| &c.rpc_secret))
+		.ok_or("No RPC secret provided")?;
+	let network_key = NetworkKey::from_slice(
+		&hex::decode(net_key_hex_str).err_context("Invalid RPC secret key (bad hex)")?[..],
+	)
+	.ok_or("Invalid RPC secret provided (wrong length)")?;
+
+	// Generate a temporary keypair for our RPC client
+	let (_pk, sk) = sodiumoxide::crypto::sign::ed25519::gen_keypair();
+
+	let netapp = NetApp::new(network_key, sk);
+
+	// Find and parse the address of the target host
+	let (id, addr) = if let Some(h) = opt.rpc_host {
+		let (id, addrs) = parse_and_resolve_peer_addr(&h).ok_or_else(|| format!("Invalid RPC remote node identifier: {}. Expected format is <pubkey>@<IP or hostname>:<port>.", h))?;
+		(id, addrs[0])
+	} else {
+		let node_id = garage_rpc::system::read_node_id(&config.as_ref().unwrap().metadata_dir)
+			.err_context(READ_KEY_ERROR)?;
+		if let Some(a) = config.as_ref().map(|c| c.rpc_public_addr).flatten() {
+			(node_id, a)
+		} else {
+			let default_addr = SocketAddr::new(
+				"127.0.0.1".parse().unwrap(),
+				config.as_ref().unwrap().rpc_bind_addr.port(),
+			);
+			warn!(
+				"Trying to contact Garage node at default address {}",
+				default_addr
+			);
+			warn!("If this doesn't work, consider adding rpc_public_addr in your config file or specifying the -h command line parameter.");
+			(node_id, default_addr)
 		}
 	};
 
-	let rpc_http_cli =
-		Arc::new(RpcHttpClient::new(8, &tls_config).expect("Could not create RPC client"));
-	let membership_rpc_cli =
-		RpcAddrClient::new(rpc_http_cli.clone(), MEMBERSHIP_RPC_PATH.to_string());
-	let admin_rpc_cli = RpcAddrClient::new(rpc_http_cli.clone(), ADMIN_RPC_PATH.to_string());
+	// Connect to target host
+	netapp.clone().try_connect(addr, id).await
+		.err_context("Unable to connect to destination RPC host. Check that you are using the same value of rpc_secret as them, and that you have their correct public key.")?;
 
-	cli_cmd(opt.cmd, membership_rpc_cli, admin_rpc_cli, opt.rpc_host).await
+	let system_rpc_endpoint = netapp.endpoint::<SystemRpc, ()>(SYSTEM_RPC_PATH.into());
+	let admin_rpc_endpoint = netapp.endpoint::<AdminRpc, ()>(ADMIN_RPC_PATH.into());
+
+	cli_command_dispatch(opt.cmd, &system_rpc_endpoint, &admin_rpc_endpoint, id).await
 }

src/garage/repair.rs

@@ -27,42 +27,38 @@ impl Repair {
 		opt: RepairOpt,
 		must_exit: watch::Receiver<bool>,
 	) -> Result<(), Error> {
-		let todo = |x| opt.what.as_ref().map(|y| *y == x).unwrap_or(true);
-
-		if todo(RepairWhat::Tables) {
-			info!("Launching a full sync of tables");
-			self.garage.bucket_table.syncer.add_full_sync();
-			self.garage.object_table.syncer.add_full_sync();
-			self.garage.version_table.syncer.add_full_sync();
-			self.garage.block_ref_table.syncer.add_full_sync();
-			self.garage.key_table.syncer.add_full_sync();
+		match opt.what {
+			RepairWhat::Tables => {
+				info!("Launching a full sync of tables");
+				self.garage.bucket_table.syncer.add_full_sync();
+				self.garage.object_table.syncer.add_full_sync();
+				self.garage.version_table.syncer.add_full_sync();
+				self.garage.block_ref_table.syncer.add_full_sync();
+				self.garage.key_table.syncer.add_full_sync();
+			}
+			RepairWhat::Versions => {
+				info!("Repairing the versions table");
+				self.repair_versions(&must_exit).await?;
+			}
+			RepairWhat::BlockRefs => {
+				info!("Repairing the block refs table");
+				self.repair_block_ref(&must_exit).await?;
+			}
+			RepairWhat::Blocks => {
+				info!("Repairing the stored blocks");
+				self.garage
+					.block_manager
+					.repair_data_store(&must_exit)
+					.await?;
+			}
+			RepairWhat::Scrub { tranquility } => {
+				info!("Verifying integrity of stored blocks");
+				self.garage
+					.block_manager
+					.scrub_data_store(&must_exit, tranquility)
+					.await?;
+			}
 		}
-
-		// TODO: wait for full sync to finish before proceeding to the rest?
-
-		if todo(RepairWhat::Versions) {
-			info!("Repairing the versions table");
-			self.repair_versions(&must_exit).await?;
-		}
-
-		if todo(RepairWhat::BlockRefs) {
-			info!("Repairing the block refs table");
-			self.repair_block_ref(&must_exit).await?;
-		}
-
-		if opt.what.is_none() {
-			info!("Repairing the RC");
-			self.repair_rc(&must_exit).await?;
-		}
-
-		if todo(RepairWhat::Blocks) {
-			info!("Repairing the stored blocks");
-			self.garage
-				.block_manager
-				.repair_data_store(&must_exit)
-				.await?;
-		}
-
 		Ok(())
 	}
 
@@ -150,10 +146,4 @@ impl Repair {
 		}
 		Ok(())
 	}
-
-	async fn repair_rc(&self, _must_exit: &watch::Receiver<bool>) -> Result<(), Error> {
-		// TODO
-		warn!("repair_rc: not implemented");
-		Ok(())
-	}
 }

src/garage/server.rs

@@ -1,7 +1,5 @@
 use std::path::PathBuf;
-use std::sync::Arc;
 
-use futures_util::future::*;
 use tokio::sync::watch;
 
 use garage_util::background::*;
@@ -10,20 +8,9 @@ use garage_util::error::Error;
 
 use garage_api::run_api_server;
 use garage_model::garage::Garage;
-use garage_rpc::rpc_server::RpcServer;
 use garage_web::run_web_server;
 
-use crate::admin_rpc::*;
-
-async fn shutdown_signal(send_cancel: watch::Sender<bool>) -> Result<(), Error> {
-	// Wait for the CTRL+C signal
-	tokio::signal::ctrl_c()
-		.await
-		.expect("failed to install CTRL+C signal handler");
-	info!("Received CTRL+C, shutting down.");
-	send_cancel.send(true)?;
-	Ok(())
-}
+use crate::admin::*;
 
 async fn wait_from(mut chan: watch::Receiver<bool>) {
 	while !*chan.borrow() {
@@ -47,52 +34,52 @@ pub async fn run_server(config_file: PathBuf) -> Result<(), Error> {
 		.open()
 		.expect("Unable to open sled DB");
 
-	info!("Initialize RPC server...");
-	let mut rpc_server = RpcServer::new(config.rpc_bind_addr, config.rpc_tls.clone());
-
 	info!("Initializing background runner...");
-	let (send_cancel, watch_cancel) = watch::channel(false);
+	let watch_cancel = netapp::util::watch_ctrl_c();
 	let (background, await_background_done) = BackgroundRunner::new(16, watch_cancel.clone());
 
 	info!("Initializing Garage main data store...");
-	let garage = Garage::new(config.clone(), db, background, &mut rpc_server);
-	let bootstrap = garage.system.clone().bootstrap(
-		config.bootstrap_peers,
-		config.consul_host,
-		config.consul_service_name,
-	);
+	let garage = Garage::new(config.clone(), db, background);
+
+	let run_system = tokio::spawn(garage.system.clone().run(watch_cancel.clone()));
 
 	info!("Crate admin RPC handler...");
-	AdminRpcHandler::new(garage.clone()).register_handler(&mut rpc_server);
+	AdminRpcHandler::new(garage.clone());
 
-	info!("Initializing RPC and API servers...");
-	let run_rpc_server = Arc::new(rpc_server).run(wait_from(watch_cancel.clone()));
-	let api_server = run_api_server(garage.clone(), wait_from(watch_cancel.clone()));
-	let web_server = run_web_server(garage, wait_from(watch_cancel.clone()));
+	info!("Initializing API server...");
+	let api_server = tokio::spawn(run_api_server(
+		garage.clone(),
+		wait_from(watch_cancel.clone()),
+	));
 
-	futures::try_join!(
-		bootstrap.map(|()| {
-			info!("Bootstrap done");
-			Ok(())
-		}),
-		run_rpc_server.map(|rv| {
-			info!("RPC server exited");
-			rv
-		}),
-		api_server.map(|rv| {
-			info!("API server exited");
-			rv
-		}),
-		web_server.map(|rv| {
-			info!("Web server exited");
-			rv
-		}),
-		await_background_done.map(|rv| {
-			info!("Background runner exited: {:?}", rv);
-			Ok(())
-		}),
-		shutdown_signal(send_cancel),
-	)?;
+	info!("Initializing web server...");
+	let web_server = tokio::spawn(run_web_server(
+		garage.clone(),
+		wait_from(watch_cancel.clone()),
+	));
+
+	// Stuff runs
+
+	// When a cancel signal is sent, stuff stops
+	if let Err(e) = api_server.await? {
+		warn!("API server exited with error: {}", e);
+	}
+	if let Err(e) = web_server.await? {
+		warn!("Web server exited with error: {}", e);
+	}
+
+	// Remove RPC handlers for system to break reference cycles
+	garage.system.netapp.drop_all_handlers();
+
+	// Await for netapp RPC system to end
+	run_system.await?;
+
+	// Break last reference cycles so that stuff can terminate properly
+	garage.break_reference_cycles();
+	drop(garage);
+
+	// Await for all background tasks to end
+	await_background_done.await?;
 
 	info!("Cleaning up...");
 

src/model/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "garage_model"
-version = "0.3.0"
+version = "0.4.0"
 authors = ["Alex Auvolat <alex@adnab.me>"]
 edition = "2018"
 license = "AGPL-3.0"
@@ -13,10 +13,11 @@ path = "lib.rs"
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [dependencies]
-garage_rpc = { version = "0.3.0", path = "../rpc" }
-garage_table = { version = "0.3.0", path = "../table" }
-garage_util = { version = "0.3.0", path = "../util" }
+garage_rpc = { version = "0.4.0", path = "../rpc" }
+garage_table = { version = "0.4.0", path = "../table" }
+garage_util = { version = "0.4.0", path = "../util" }
 
+async-trait = "0.1.7"
 arc-swap = "1.0"
 hex = "0.4"
 log = "0.4"
@@ -31,3 +32,6 @@ serde_bytes = "0.11"
 futures = "0.3"
 futures-util = "0.3"
 tokio = { version = "1.0", default-features = false, features = ["rt", "rt-multi-thread", "io-util", "net", "time", "macros", "sync", "signal", "fs"] }
+
+#netapp = { version = "0.3.0", git = "https://git.deuxfleurs.fr/lx/netapp" }
+netapp = "0.3.0"

src/model/block.rs

@@ -3,6 +3,7 @@ use std::sync::Arc;
 use std::time::Duration;
 
 use arc_swap::ArcSwapOption;
+use async_trait::async_trait;
 use futures::future::*;
 use futures::select;
 use serde::{Deserialize, Serialize};
@@ -11,12 +12,12 @@ use tokio::io::{AsyncReadExt, AsyncWriteExt};
 use tokio::sync::{watch, Mutex, Notify};
 
 use garage_util::data::*;
-use garage_util::error::Error;
+use garage_util::error::*;
 use garage_util::time::*;
+use garage_util::tranquilizer::Tranquilizer;
 
-use garage_rpc::membership::System;
-use garage_rpc::rpc_client::*;
-use garage_rpc::rpc_server::*;
+use garage_rpc::system::System;
+use garage_rpc::*;
 
 use garage_table::replication::{TableReplication, TableShardedReplication};
 
@@ -28,6 +29,7 @@ use crate::garage::Garage;
 pub const INLINE_THRESHOLD: usize = 3072;
 
 pub const BACKGROUND_WORKERS: u64 = 1;
+pub const BACKGROUND_TRANQUILITY: u32 = 3;
 
 const BLOCK_RW_TIMEOUT: Duration = Duration::from_secs(42);
 const BLOCK_GC_TIMEOUT: Duration = Duration::from_secs(60);
@@ -36,7 +38,7 @@ const RESYNC_RETRY_TIMEOUT: Duration = Duration::from_secs(10);
 
 /// RPC messages used to share blocks of data between nodes
 #[derive(Debug, Serialize, Deserialize)]
-pub enum Message {
+pub enum BlockRpc {
 	Ok,
 	/// Message to ask for a block of data, by hash
 	GetBlock(Hash),
@@ -60,7 +62,9 @@ pub struct PutBlockMessage {
 	pub data: Vec<u8>,
 }
 
-impl RpcMessage for Message {}
+impl Rpc for BlockRpc {
+	type Response = Result<BlockRpc, Error>;
+}
 
 /// The block manager, handling block exchange between nodes, and block storage on local node
 pub struct BlockManager {
@@ -68,8 +72,8 @@ pub struct BlockManager {
 	pub replication: TableShardedReplication,
 	/// Directory in which block are stored
 	pub data_dir: PathBuf,
-	/// Lock to prevent concurrent edition of the directory
-	pub data_dir_lock: Mutex<()>,
+
+	mutation_lock: Mutex<BlockManagerLocked>,
 
 	rc: sled::Tree,
 
@@ -77,17 +81,21 @@ pub struct BlockManager {
 	resync_notify: Notify,
 
 	system: Arc<System>,
-	rpc_client: Arc<RpcClient<Message>>,
+	endpoint: Arc<Endpoint<BlockRpc, Self>>,
 	pub(crate) garage: ArcSwapOption<Garage>,
 }
 
+// This custom struct contains functions that must only be ran
+// when the lock is held. We ensure that it is the case by storing
+// it INSIDE a Mutex.
+struct BlockManagerLocked();
+
 impl BlockManager {
 	pub fn new(
 		db: &sled::Db,
 		data_dir: PathBuf,
 		replication: TableShardedReplication,
 		system: Arc<System>,
-		rpc_server: &mut RpcServer,
 	) -> Arc<Self> {
 		let rc = db
 			.open_tree("block_local_rc")
@@ -97,144 +105,143 @@ impl BlockManager {
 			.open_tree("block_local_resync_queue")
 			.expect("Unable to open block_local_resync_queue tree");
 
-		let rpc_path = "block_manager";
-		let rpc_client = system.rpc_client::<Message>(rpc_path);
+		let endpoint = system
+			.netapp
+			.endpoint("garage_model/block.rs/Rpc".to_string());
+
+		let manager_locked = BlockManagerLocked();
 
 		let block_manager = Arc::new(Self {
 			replication,
 			data_dir,
-			data_dir_lock: Mutex::new(()),
+			mutation_lock: Mutex::new(manager_locked),
 			rc,
 			resync_queue,
 			resync_notify: Notify::new(),
 			system,
-			rpc_client,
+			endpoint,
 			garage: ArcSwapOption::from(None),
 		});
-		block_manager
-			.clone()
-			.register_handler(rpc_server, rpc_path.into());
+		block_manager.endpoint.set_handler(block_manager.clone());
+
 		block_manager
 	}
 
-	fn register_handler(self: Arc<Self>, rpc_server: &mut RpcServer, path: String) {
-		let self2 = self.clone();
-		rpc_server.add_handler::<Message, _, _>(path, move |msg, _addr| {
-			let self2 = self2.clone();
-			async move { self2.handle(&msg).await }
-		});
+	// ---- Public interface ----
 
-		let self2 = self.clone();
-		self.rpc_client
-			.set_local_handler(self.system.id, move |msg| {
-				let self2 = self2.clone();
-				async move { self2.handle(&msg).await }
-			});
-	}
+	/// Ask nodes that might have a block for it
+	pub async fn rpc_get_block(&self, hash: &Hash) -> Result<Vec<u8>, Error> {
+		let who = self.replication.read_nodes(hash);
+		let resps = self
+			.system
+			.rpc
+			.try_call_many(
+				&self.endpoint,
+				&who[..],
+				BlockRpc::GetBlock(*hash),
+				RequestStrategy::with_priority(PRIO_NORMAL)
+					.with_quorum(1)
+					.with_timeout(BLOCK_RW_TIMEOUT)
+					.interrupt_after_quorum(true),
+			)
+			.await?;
 
-	async fn handle(self: Arc<Self>, msg: &Message) -> Result<Message, Error> {
-		match msg {
-			Message::PutBlock(m) => self.write_block(&m.hash, &m.data).await,
-			Message::GetBlock(h) => self.read_block(h).await,
-			Message::NeedBlockQuery(h) => self.need_block(h).await.map(Message::NeedBlockReply),
-			_ => Err(Error::BadRpc("Unexpected RPC message".to_string())),
-		}
-	}
-
-	pub fn spawn_background_worker(self: Arc<Self>) {
-		// Launch 2 simultaneous workers for background resync loop preprocessing <= TODO actually this
-		// launches only one worker with current value of BACKGROUND_WORKERS
-		for i in 0..BACKGROUND_WORKERS {
-			let bm2 = self.clone();
-			let background = self.system.background.clone();
-			tokio::spawn(async move {
-				tokio::time::sleep(Duration::from_secs(10 * (i + 1))).await;
-				background.spawn_worker(format!("block resync worker {}", i), move |must_exit| {
-					bm2.resync_loop(must_exit)
-				});
-			});
-		}
-	}
-
-	/// Write a block to disk
-	async fn write_block(&self, hash: &Hash, data: &[u8]) -> Result<Message, Error> {
-		let _lock = self.data_dir_lock.lock().await;
-
-		let mut path = self.block_dir(hash);
-		fs::create_dir_all(&path).await?;
-
-		path.push(hex::encode(hash));
-		if fs::metadata(&path).await.is_ok() {
-			return Ok(Message::Ok);
-		}
-
-		let mut f = fs::File::create(path).await?;
-		f.write_all(data).await?;
-		drop(f);
-
-		Ok(Message::Ok)
-	}
-
-	/// Read block from disk, verifying it's integrity
-	async fn read_block(&self, hash: &Hash) -> Result<Message, Error> {
-		let path = self.block_path(hash);
-
-		let mut f = match fs::File::open(&path).await {
-			Ok(f) => f,
-			Err(e) => {
-				// Not found but maybe we should have had it ??
-				self.put_to_resync(hash, Duration::from_millis(0))?;
-				return Err(Into::into(e));
+		for resp in resps {
+			if let BlockRpc::PutBlock(msg) = resp {
+				return Ok(msg.data);
 			}
-		};
-		let mut data = vec![];
-		f.read_to_end(&mut data).await?;
-		drop(f);
+		}
+		Err(Error::Message(format!(
+			"Unable to read block {:?}: no valid blocks returned",
+			hash
+		)))
+	}
 
-		if blake2sum(&data[..]) != *hash {
-			let _lock = self.data_dir_lock.lock().await;
-			warn!(
-				"Block {:?} is corrupted. Renaming to .corrupted and resyncing.",
-				hash
-			);
-			let mut path2 = path.clone();
-			path2.set_extension(".corrupted");
-			fs::rename(path, path2).await?;
-			self.put_to_resync(&hash, Duration::from_millis(0))?;
-			return Err(Error::CorruptData(*hash));
+	/// Send block to nodes that should have it
+	pub async fn rpc_put_block(&self, hash: Hash, data: Vec<u8>) -> Result<(), Error> {
+		let who = self.replication.write_nodes(&hash);
+		self.system
+			.rpc
+			.try_call_many(
+				&self.endpoint,
+				&who[..],
+				BlockRpc::PutBlock(PutBlockMessage { hash, data }),
+				RequestStrategy::with_priority(PRIO_NORMAL)
+					.with_quorum(self.replication.write_quorum())
+					.with_timeout(BLOCK_RW_TIMEOUT),
+			)
+			.await?;
+		Ok(())
+	}
+
+	/// Launch the repair procedure on the data store
+	///
+	/// This will list all blocks locally present, as well as those
+	/// that are required because of refcount > 0, and will try
+	/// to fix any mismatch between the two.
+	pub async fn repair_data_store(&self, must_exit: &watch::Receiver<bool>) -> Result<(), Error> {
+		// 1. Repair blocks from RC table
+		let garage = self.garage.load_full().unwrap();
+		let mut last_hash = None;
+		for (i, entry) in garage.block_ref_table.data.store.iter().enumerate() {
+			let (_k, v_bytes) = entry?;
+			let block_ref = rmp_serde::decode::from_read_ref::<_, BlockRef>(v_bytes.as_ref())?;
+			if Some(&block_ref.block) == last_hash.as_ref() {
+				continue;
+			}
+			if !block_ref.deleted.get() {
+				last_hash = Some(block_ref.block);
+				self.put_to_resync(&block_ref.block, Duration::from_secs(0))?;
+			}
+			if i & 0xFF == 0 && *must_exit.borrow() {
+				return Ok(());
+			}
 		}
 
-		Ok(Message::PutBlock(PutBlockMessage { hash: *hash, data }))
+		// 2. Repair blocks actually on disk
+		// Lists all blocks on disk and adds them to the resync queue.
+		// This allows us to find blocks we are storing but don't actually need,
+		// so that we can offload them if necessary and then delete them locally.
+		self.for_each_file(
+			(),
+			move |_, hash| async move { self.put_to_resync(&hash, Duration::from_secs(0)) },
+			must_exit,
+		)
+		.await
 	}
 
-	/// Check if this node should have a block, but don't actually have it
-	async fn need_block(&self, hash: &Hash) -> Result<bool, Error> {
-		let needed = self
-			.rc
-			.get(hash.as_ref())?
-			.map(|x| u64_from_be_bytes(x) > 0)
-			.unwrap_or(false);
-		if needed {
-			let path = self.block_path(hash);
-			let exists = fs::metadata(&path).await.is_ok();
-			Ok(!exists)
-		} else {
-			Ok(false)
-		}
+	/// Verify integrity of each block on disk. Use `speed_limit` to limit the load generated by
+	/// this function.
+	pub async fn scrub_data_store(
+		&self,
+		must_exit: &watch::Receiver<bool>,
+		tranquility: u32,
+	) -> Result<(), Error> {
+		let tranquilizer = Tranquilizer::new(30);
+		self.for_each_file(
+			tranquilizer,
+			move |mut tranquilizer, hash| async move {
+				let _ = self.read_block(&hash).await;
+				tranquilizer.tranquilize(tranquility).await;
+				Ok(tranquilizer)
+			},
+			must_exit,
+		)
+		.await
 	}
 
-	fn block_dir(&self, hash: &Hash) -> PathBuf {
-		let mut path = self.data_dir.clone();
-		path.push(hex::encode(&hash.as_slice()[0..1]));
-		path.push(hex::encode(&hash.as_slice()[1..2]));
-		path
+	/// Get lenght of resync queue
+	pub fn resync_queue_len(&self) -> usize {
+		self.resync_queue.len()
 	}
-	fn block_path(&self, hash: &Hash) -> PathBuf {
-		let mut path = self.block_dir(hash);
-		path.push(hex::encode(hash.as_ref()));
-		path
+
+	/// Get number of items in the refcount table
+	pub fn rc_len(&self) -> usize {
+		self.rc.len()
 	}
 
+	//// ----- Managing the reference counter ----
+
 	/// Increment the number of time a block is used, putting it to resynchronization if it is
 	/// required, but not known
 	pub fn block_incref(&self, hash: &Hash) -> Result<(), Error> {
@@ -244,7 +251,7 @@ impl BlockManager {
 		})?;
 		let old_rc = old_rc.map(u64_from_be_bytes).unwrap_or(0);
 		if old_rc == 0 {
-			self.put_to_resync(&hash, BLOCK_RW_TIMEOUT)?;
+			self.put_to_resync(hash, BLOCK_RW_TIMEOUT)?;
 		}
 		Ok(())
 	}
@@ -260,11 +267,101 @@ impl BlockManager {
 			}
 		})?;
 		if new_rc.is_none() {
-			self.put_to_resync(&hash, BLOCK_GC_TIMEOUT)?;
+			self.put_to_resync(hash, BLOCK_GC_TIMEOUT)?;
 		}
 		Ok(())
 	}
 
+	/// Read a block's reference count
+	pub fn get_block_rc(&self, hash: &Hash) -> Result<u64, Error> {
+		Ok(self
+			.rc
+			.get(hash.as_ref())?
+			.map(u64_from_be_bytes)
+			.unwrap_or(0))
+	}
+
+	// ---- Reading and writing blocks locally ----
+
+	/// Write a block to disk
+	async fn write_block(&self, hash: &Hash, data: &[u8]) -> Result<BlockRpc, Error> {
+		self.mutation_lock
+			.lock()
+			.await
+			.write_block(hash, data, self)
+			.await
+	}
+
+	/// Read block from disk, verifying it's integrity
+	async fn read_block(&self, hash: &Hash) -> Result<BlockRpc, Error> {
+		let path = self.block_path(hash);
+
+		let mut f = match fs::File::open(&path).await {
+			Ok(f) => f,
+			Err(e) => {
+				// Not found but maybe we should have had it ??
+				self.put_to_resync(hash, Duration::from_millis(0))?;
+				return Err(Into::into(e));
+			}
+		};
+		let mut data = vec![];
+		f.read_to_end(&mut data).await?;
+		drop(f);
+
+		if blake2sum(&data[..]) != *hash {
+			self.mutation_lock
+				.lock()
+				.await
+				.move_block_to_corrupted(hash, self)
+				.await?;
+			return Err(Error::CorruptData(*hash));
+		}
+
+		Ok(BlockRpc::PutBlock(PutBlockMessage { hash: *hash, data }))
+	}
+
+	/// Check if this node should have a block, but don't actually have it
+	async fn need_block(&self, hash: &Hash) -> Result<bool, Error> {
+		let BlockStatus { exists, needed } = self
+			.mutation_lock
+			.lock()
+			.await
+			.check_block_status(hash, self)
+			.await?;
+		Ok(needed && !exists)
+	}
+
+	/// Utility: gives the path of the directory in which a block should be found
+	fn block_dir(&self, hash: &Hash) -> PathBuf {
+		let mut path = self.data_dir.clone();
+		path.push(hex::encode(&hash.as_slice()[0..1]));
+		path.push(hex::encode(&hash.as_slice()[1..2]));
+		path
+	}
+
+	/// Utility: give the full path where a block should be found
+	fn block_path(&self, hash: &Hash) -> PathBuf {
+		let mut path = self.block_dir(hash);
+		path.push(hex::encode(hash.as_ref()));
+		path
+	}
+
+	// ---- Resync loop ----
+
+	pub fn spawn_background_worker(self: Arc<Self>) {
+		// Launch 2 simultaneous workers for background resync loop preprocessing
+		for i in 0..BACKGROUND_WORKERS {
+			let bm2 = self.clone();
+			let background = self.system.background.clone();
+			tokio::spawn(async move {
+				tokio::time::sleep(Duration::from_secs(10 * (i + 1))).await;
+				background.spawn_worker(format!("block resync worker {}", i), move |must_exit| {
+					bm2.resync_loop(must_exit)
+				});
+			});
+		}
+	}
+
 	fn put_to_resync(&self, hash: &Hash, delay: Duration) -> Result<(), Error> {
 		let when = now_msec() + delay.as_millis() as u64;
 		trace!("Put resync_queue: {} {:?}", when, hash);
@@ -276,20 +373,33 @@ impl BlockManager {
 	}
 
 	async fn resync_loop(self: Arc<Self>, mut must_exit: watch::Receiver<bool>) {
+		let mut tranquilizer = Tranquilizer::new(30);
+
 		while !*must_exit.borrow() {
-			if let Err(e) = self.resync_iter(&mut must_exit).await {
-				warn!("Error in block resync loop: {}", e);
-				select! {
-					_ = tokio::time::sleep(Duration::from_secs(1)).fuse() => {},
-					_ = must_exit.changed().fuse() => {},
+			match self.resync_iter(&mut must_exit).await {
+				Ok(true) => {
+					tranquilizer.tranquilize(BACKGROUND_TRANQUILITY).await;
+				}
+				Ok(false) => {
+					tranquilizer.reset();
+				}
+				Err(e) => {
+					// The errors that we have here are only Sled errors
+					// We don't really know how to handle them so just ¯\_(ツ)_/¯
+					// (there is kind of an assumption that Sled won't error on us,
+					// if it does there is not much we can do -- TODO should we just panic?)
+					error!(
+						"Could not do a resync iteration: {} (this is a very bad error)",
+						e
+					);
+					tranquilizer.reset();
 				}
 			}
 		}
 	}
 
-	async fn resync_iter(&self, must_exit: &mut watch::Receiver<bool>) -> Result<(), Error> {
-		if let Some(first_item) = self.resync_queue.iter().next() {
-			let (time_bytes, hash_bytes) = first_item?;
+	async fn resync_iter(&self, must_exit: &mut watch::Receiver<bool>) -> Result<bool, Error> {
+		if let Some((time_bytes, hash_bytes)) = self.resync_queue.pop_min()? {
 			let time_msec = u64_from_be_bytes(&time_bytes[0..8]);
 			let now = now_msec();
 			if now >= time_msec {
@@ -299,36 +409,33 @@ impl BlockManager {
 					warn!("Error when resyncing {:?}: {}", hash, e);
 					self.put_to_resync(&hash, RESYNC_RETRY_TIMEOUT)?;
 				}
-				self.resync_queue.remove(&time_bytes)?;
-				res?; // propagate error to delay main loop
+				Ok(true)
 			} else {
+				self.resync_queue.insert(time_bytes, hash_bytes)?;
 				let delay = tokio::time::sleep(Duration::from_millis(time_msec - now));
 				select! {
 					_ = delay.fuse() => {},
 					_ = self.resync_notify.notified().fuse() => {},
 					_ = must_exit.changed().fuse() => {},
 				}
+				Ok(false)
 			}
 		} else {
 			select! {
 				_ = self.resync_notify.notified().fuse() => {},
 				_ = must_exit.changed().fuse() => {},
 			}
+			Ok(false)
 		}
-		Ok(())
 	}
 
 	async fn resync_block(&self, hash: &Hash) -> Result<(), Error> {
-		let lock = self.data_dir_lock.lock().await;
-
-		let path = self.block_path(hash);
-
-		let exists = fs::metadata(&path).await.is_ok();
-		let needed = self
-			.rc
-			.get(hash.as_ref())?
-			.map(|x| u64_from_be_bytes(x) > 0)
-			.unwrap_or(false);
+		let BlockStatus { exists, needed } = self
+			.mutation_lock
+			.lock()
+			.await
+			.check_block_status(hash, self)
+			.await?;
 
 		if exists != needed {
 			info!(
@@ -340,23 +447,28 @@ impl BlockManager {
 		if exists && !needed {
 			trace!("Offloading block {:?}", hash);
 
-			let mut who = self.replication.write_nodes(&hash);
+			let mut who = self.replication.write_nodes(hash);
 			if who.len() < self.replication.write_quorum() {
 				return Err(Error::Message("Not trying to offload block because we don't have a quorum of nodes to write to".to_string()));
 			}
 			who.retain(|id| *id != self.system.id);
 
-			let msg = Arc::new(Message::NeedBlockQuery(*hash));
+			let msg = Arc::new(BlockRpc::NeedBlockQuery(*hash));
 			let who_needs_fut = who.iter().map(|to| {
-				self.rpc_client
-					.call_arc(*to, msg.clone(), NEED_BLOCK_QUERY_TIMEOUT)
+				self.system.rpc.call_arc(
+					&self.endpoint,
+					*to,
+					msg.clone(),
+					RequestStrategy::with_priority(PRIO_BACKGROUND)
+						.with_timeout(NEED_BLOCK_QUERY_TIMEOUT),
+				)
 			});
 			let who_needs_resps = join_all(who_needs_fut).await;
 
 			let mut need_nodes = vec![];
 			for (node, needed) in who.iter().zip(who_needs_resps.into_iter()) {
-				match needed? {
-					Message::NeedBlockReply(needed) => {
+				match needed.err_context("NeedBlockQuery RPC")? {
+					BlockRpc::NeedBlockReply(needed) => {
 						if needed {
 							need_nodes.push(*node);
 						}
@@ -376,12 +488,15 @@ impl BlockManager {
 					need_nodes.len()
 				);
 
-				let put_block_message = self.read_block(hash).await?;
-				self.rpc_client
+				let put_block_message = self.read_block(hash).await.err_context("PutBlock RPC")?;
+				self.system
+					.rpc
 					.try_call_many(
+						&self.endpoint,
 						&need_nodes[..],
 						put_block_message,
-						RequestStrategy::with_quorum(need_nodes.len())
+						RequestStrategy::with_priority(PRIO_BACKGROUND)
+							.with_quorum(need_nodes.len())
 							.with_timeout(BLOCK_RW_TIMEOUT),
 					)
 					.await?;
@@ -393,141 +508,169 @@ impl BlockManager {
 				who.len()
 			);
 
-			fs::remove_file(path).await?;
+			self.mutation_lock
+				.lock()
+				.await
+				.delete_if_unneeded(hash, self)
+				.await?;
 		}
 
 		if needed && !exists {
-			drop(lock);
-
 			// TODO find a way to not do this if they are sending it to us
 			// Let's suppose this isn't an issue for now with the BLOCK_RW_TIMEOUT delay
 			// between the RC being incremented and this part being called.
-			let block_data = self.rpc_get_block(&hash).await?;
+			let block_data = self.rpc_get_block(hash).await?;
 			self.write_block(hash, &block_data[..]).await?;
 		}
 
 		Ok(())
 	}
 
-	/// Ask nodes that might have a block for it
-	pub async fn rpc_get_block(&self, hash: &Hash) -> Result<Vec<u8>, Error> {
-		let who = self.replication.read_nodes(&hash);
-		let resps = self
-			.rpc_client
-			.try_call_many(
-				&who[..],
-				Message::GetBlock(*hash),
-				RequestStrategy::with_quorum(1)
-					.with_timeout(BLOCK_RW_TIMEOUT)
-					.interrupt_after_quorum(true),
-			)
-			.await?;
-
-		for resp in resps {
-			if let Message::PutBlock(msg) = resp {
-				return Ok(msg.data);
-			}
-		}
-		Err(Error::Message(format!(
-			"Unable to read block {:?}: no valid blocks returned",
-			hash
-		)))
+	async fn for_each_file<F, Fut, State>(
+		&self,
+		state: State,
+		mut f: F,
+		must_exit: &watch::Receiver<bool>,
+	) -> Result<(), Error>
+	where
+		F: FnMut(State, Hash) -> Fut + Send,
+		Fut: Future<Output = Result<State, Error>> + Send,
+		State: Send,
+	{
+		self.for_each_file_rec(&self.data_dir, state, &mut f, must_exit)
+			.await
+			.map(|_| ())
 	}
 
-	/// Send block to nodes that should have it
-	pub async fn rpc_put_block(&self, hash: Hash, data: Vec<u8>) -> Result<(), Error> {
-		let who = self.replication.write_nodes(&hash);
-		self.rpc_client
-			.try_call_many(
-				&who[..],
-				Message::PutBlock(PutBlockMessage { hash, data }),
-				RequestStrategy::with_quorum(self.replication.write_quorum())
-					.with_timeout(BLOCK_RW_TIMEOUT),
-			)
-			.await?;
-		Ok(())
-	}
-
-	pub async fn repair_data_store(&self, must_exit: &watch::Receiver<bool>) -> Result<(), Error> {
-		// 1. Repair blocks from RC table
-		let garage = self.garage.load_full().unwrap();
-		let mut last_hash = None;
-		let mut i = 0usize;
-		for entry in garage.block_ref_table.data.store.iter() {
-			let (_k, v_bytes) = entry?;
-			let block_ref = rmp_serde::decode::from_read_ref::<_, BlockRef>(v_bytes.as_ref())?;
-			if Some(&block_ref.block) == last_hash.as_ref() {
-				continue;
-			}
-			if !block_ref.deleted.get() {
-				last_hash = Some(block_ref.block);
-				self.put_to_resync(&block_ref.block, Duration::from_secs(0))?;
-			}
-			i += 1;
-			if i & 0xFF == 0 && *must_exit.borrow() {
-				return Ok(());
-			}
-		}
-
-		// 2. Repair blocks actually on disk
-		self.repair_aux_read_dir_rec(&self.data_dir, must_exit)
-			.await?;
-
-		Ok(())
-	}
-
-	fn repair_aux_read_dir_rec<'a>(
+	fn for_each_file_rec<'a, F, Fut, State>(
 		&'a self,
 		path: &'a Path,
+		mut state: State,
+		f: &'a mut F,
 		must_exit: &'a watch::Receiver<bool>,
-	) -> BoxFuture<'a, Result<(), Error>> {
-		// Lists all blocks on disk and adds them to the resync queue.
-		// This allows us to find blocks we are storing but don't actually need,
-		// so that we can offload them if necessary and then delete them locally.
+	) -> BoxFuture<'a, Result<State, Error>>
+	where
+		F: FnMut(State, Hash) -> Fut + Send,
+		Fut: Future<Output = Result<State, Error>> + Send,
+		State: Send + 'a,
+	{
 		async move {
 			let mut ls_data_dir = fs::read_dir(path).await?;
-			loop {
-				let data_dir_ent = ls_data_dir.next_entry().await?;
-				let data_dir_ent = match data_dir_ent {
-					Some(x) => x,
-					None => break,
-				};
+			while let Some(data_dir_ent) = ls_data_dir.next_entry().await? {
+				if *must_exit.borrow() {
+					break;
+				}
+
 				let name = data_dir_ent.file_name();
-				let name = match name.into_string() {
-					Ok(x) => x,
-					Err(_) => continue,
+				let name = if let Ok(n) = name.into_string() {
+					n
+				} else {
+					continue;
 				};
 				let ent_type = data_dir_ent.file_type().await?;
 
 				if name.len() == 2 && hex::decode(&name).is_ok() && ent_type.is_dir() {
-					self.repair_aux_read_dir_rec(&data_dir_ent.path(), must_exit)
+					state = self
+						.for_each_file_rec(&data_dir_ent.path(), state, f, must_exit)
 						.await?;
 				} else if name.len() == 64 {
-					let hash_bytes = match hex::decode(&name) {
-						Ok(h) => h,
-						Err(_) => continue,
+					let hash_bytes = if let Ok(h) = hex::decode(&name) {
+						h
+					} else {
+						continue;
 					};
 					let mut hash = [0u8; 32];
 					hash.copy_from_slice(&hash_bytes[..]);
-					self.put_to_resync(&hash.into(), Duration::from_secs(0))?;
-				}
-
-				if *must_exit.borrow() {
-					break;
+					state = f(state, hash.into()).await?;
 				}
 			}
-			Ok(())
+			Ok(state)
 		}
 		.boxed()
 	}
+}
 
-	/// Get lenght of resync queue
-	pub fn resync_queue_len(&self) -> usize {
-		self.resync_queue.len()
+#[async_trait]
+impl EndpointHandler<BlockRpc> for BlockManager {
+	async fn handle(
+		self: &Arc<Self>,
+		message: &BlockRpc,
+		_from: NodeID,
+	) -> Result<BlockRpc, Error> {
+		match message {
+			BlockRpc::PutBlock(m) => self.write_block(&m.hash, &m.data).await,
+			BlockRpc::GetBlock(h) => self.read_block(h).await,
+			BlockRpc::NeedBlockQuery(h) => self.need_block(h).await.map(BlockRpc::NeedBlockReply),
+			_ => Err(Error::BadRpc("Unexpected RPC message".to_string())),
+		}
+	}
+}
+
+struct BlockStatus {
+	exists: bool,
+	needed: bool,
+}
+
+impl BlockManagerLocked {
+	async fn check_block_status(
+		&self,
+		hash: &Hash,
+		mgr: &BlockManager,
+	) -> Result<BlockStatus, Error> {
+		let path = mgr.block_path(hash);
+
+		let exists = fs::metadata(&path).await.is_ok();
+		let needed = mgr.get_block_rc(hash)? > 0;
+
+		Ok(BlockStatus { exists, needed })
 	}
 
-	pub fn rc_len(&self) -> usize {
-		self.rc.len()
+	async fn write_block(
+		&self,
+		hash: &Hash,
+		data: &[u8],
+		mgr: &BlockManager,
+	) -> Result<BlockRpc, Error> {
+		let mut path = mgr.block_dir(hash);
+		fs::create_dir_all(&path).await?;
+
+		path.push(hex::encode(hash));
+		if fs::metadata(&path).await.is_ok() {
+			return Ok(BlockRpc::Ok);
+		}
+
+		let mut path2 = path.clone();
+		path2.set_extension("tmp");
+		let mut f = fs::File::create(&path2).await?;
+		f.write_all(data).await?;
+		drop(f);
+
+		fs::rename(path2, path).await?;
+
+		Ok(BlockRpc::Ok)
+	}
+
+	async fn move_block_to_corrupted(&self, hash: &Hash, mgr: &BlockManager) -> Result<(), Error> {
+		warn!(
+			"Block {:?} is corrupted. Renaming to .corrupted and resyncing.",
+			hash
+		);
+		let path = mgr.block_path(hash);
+		let mut path2 = path.clone();
+		path2.set_extension("corrupted");
+		fs::rename(path, path2).await?;
+		mgr.put_to_resync(hash, Duration::from_millis(0))?;
+		Ok(())
+	}
+
+	async fn delete_if_unneeded(&self, hash: &Hash, mgr: &BlockManager) -> Result<(), Error> {
+		let BlockStatus { exists, needed } = self.check_block_status(hash, mgr).await?;
+
+		if exists && !needed {
+			let path = mgr.block_path(hash);
+			fs::remove_file(path).await?;
+		}
+		Ok(())
 	}
 }
 

src/model/garage.rs

@@ -1,11 +1,11 @@
 use std::sync::Arc;
 
+use netapp::NetworkKey;
+
 use garage_util::background::*;
 use garage_util::config::*;
 
-use garage_rpc::membership::System;
-use garage_rpc::rpc_client::RpcHttpClient;
-use garage_rpc::rpc_server::RpcServer;
+use garage_rpc::system::System;
 
 use garage_table::replication::ReplicationMode;
 use garage_table::replication::TableFullReplication;
@@ -45,26 +45,21 @@ pub struct Garage {
 
 impl Garage {
 	/// Create and run garage
-	pub fn new(
-		config: Config,
-		db: sled::Db,
-		background: Arc<BackgroundRunner>,
-		rpc_server: &mut RpcServer,
-	) -> Arc<Self> {
+	pub fn new(config: Config, db: sled::Db, background: Arc<BackgroundRunner>) -> Arc<Self> {
+		let network_key = NetworkKey::from_slice(
+			&hex::decode(&config.rpc_secret).expect("Invalid RPC secret key")[..],
+		)
+		.expect("Invalid RPC secret key");
+
 		let replication_mode = ReplicationMode::parse(&config.replication_mode)
 			.expect("Invalid replication_mode in config file.");
 
 		info!("Initialize membership management system...");
-		let rpc_http_client = Arc::new(
-			RpcHttpClient::new(config.max_concurrent_rpc_requests, &config.rpc_tls)
-				.expect("Could not create RPC client"),
-		);
 		let system = System::new(
-			config.metadata_dir.clone(),
-			rpc_http_client,
+			network_key,
 			background.clone(),
-			rpc_server,
 			replication_mode.replication_factor(),
+			&config,
 		);
 
 		let data_rep_param = TableShardedReplication {
@@ -87,13 +82,8 @@ impl Garage {
 		};
 
 		info!("Initialize block manager...");
-		let block_manager = BlockManager::new(
-			&db,
-			config.data_dir.clone(),
-			data_rep_param,
-			system.clone(),
-			rpc_server,
-		);
+		let block_manager =
+			BlockManager::new(&db, config.data_dir.clone(), data_rep_param, system.clone());
 
 		info!("Initialize block_ref_table...");
 		let block_ref_table = Table::new(
@@ -104,7 +94,6 @@ impl Garage {
 			system.clone(),
 			&db,
 			"block_ref".to_string(),
-			rpc_server,
 		);
 
 		info!("Initialize version_table...");
@@ -117,7 +106,6 @@ impl Garage {
 			system.clone(),
 			&db,
 			"version".to_string(),
-			rpc_server,
 		);
 
 		info!("Initialize object_table...");
@@ -130,7 +118,6 @@ impl Garage {
 			system.clone(),
 			&db,
 			"object".to_string(),
-			rpc_server,
 		);
 
 		info!("Initialize bucket_table...");
@@ -140,7 +127,6 @@ impl Garage {
 			system.clone(),
 			&db,
 			"bucket".to_string(),
-			rpc_server,
 		);
 
 		info!("Initialize key_table_table...");
@@ -150,7 +136,6 @@ impl Garage {
 			system.clone(),
 			&db,
 			"key".to_string(),
-			rpc_server,
 		);
 
 		info!("Initialize Garage...");
@@ -173,4 +158,9 @@ impl Garage {
 
 		garage
 	}
+
+	/// Use this for shutdown
+	pub fn break_reference_cycles(&self) {
+		self.block_manager.garage.swap(None);
+	}
 }

src/rpc/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "garage_rpc"
-version = "0.3.0"
+version = "0.4.0"
 authors = ["Alex Auvolat <alex@adnab.me>"]
 edition = "2018"
 license = "AGPL-3.0"
@@ -13,16 +13,17 @@ path = "lib.rs"
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [dependencies]
-garage_util = { version = "0.3.0", path = "../util" }
-
-garage_rpc_021 = { package = "garage_rpc", version = "0.2.1" }
+garage_util = { version = "0.4.0", path = "../util" }
 
 arc-swap = "1.0"
 bytes = "1.0"
 gethostname = "0.2"
 hex = "0.4"
 log = "0.4"
+rand = "0.8"
+sodiumoxide = { version = "0.2.5-0", package = "kuska-sodiumoxide" }
 
+async-trait = "0.1.7"
 rmp-serde = "0.15"
 serde = { version = "1.0", default-features = false, features = ["derive", "rc"] }
 serde_json = "1.0"
@@ -32,11 +33,7 @@ futures-util = "0.3"
 tokio = { version = "1.0", default-features = false, features = ["rt", "rt-multi-thread", "io-util", "net", "time", "macros", "sync", "signal", "fs"] }
 tokio-stream = { version = "0.1", features = ["net"] }
 
-http = "0.2"
-hyper = { version = "0.14", features = ["full"] }
-hyper-rustls = { version = "0.22", default-features = false }
-rustls = "0.19"
-tokio-rustls = "0.22"
-webpki = "0.21"
-
+#netapp = { version = "0.3.0", git = "https://git.deuxfleurs.fr/lx/netapp" }
+netapp = "0.3.0"
+hyper = { version = "0.14", features = ["client", "http1", "runtime", "tcp"] }
 

src/rpc/consul.rs

@@ -1,24 +1,31 @@
+use std::collections::HashMap;
 use std::net::{IpAddr, SocketAddr};
 
 use hyper::client::Client;
 use hyper::StatusCode;
 use hyper::{Body, Method, Request};
-use serde::Deserialize;
+use serde::{Deserialize, Serialize};
+
+use netapp::NodeID;
 
 use garage_util::error::Error;
 
-#[derive(Deserialize, Clone)]
-struct ConsulEntry {
-	#[serde(alias = "Address")]
+// ---- READING FROM CONSUL CATALOG ----
+
+#[derive(Deserialize, Clone, Debug)]
+struct ConsulQueryEntry {
+	#[serde(rename = "Address")]
 	address: String,
-	#[serde(alias = "ServicePort")]
+	#[serde(rename = "ServicePort")]
 	service_port: u16,
+	#[serde(rename = "NodeMeta")]
+	node_meta: HashMap<String, String>,
 }
 
 pub async fn get_consul_nodes(
 	consul_host: &str,
 	consul_service_name: &str,
-) -> Result<Vec<SocketAddr>, Error> {
+) -> Result<Vec<(NodeID, SocketAddr)>, Error> {
 	let url = format!(
 		"http://{}/v1/catalog/service/{}",
 		consul_host, consul_service_name
@@ -36,17 +43,111 @@ pub async fn get_consul_nodes(
 	}
 
 	let body = hyper::body::to_bytes(resp.into_body()).await?;
-	let entries = serde_json::from_slice::<Vec<ConsulEntry>>(body.as_ref())?;
+	let entries = serde_json::from_slice::<Vec<ConsulQueryEntry>>(body.as_ref())?;
 
 	let mut ret = vec![];
 	for ent in entries {
-		let ip = ent
-			.address
-			.parse::<IpAddr>()
-			.map_err(|e| Error::Message(format!("Could not parse IP address: {}", e)))?;
-		ret.push(SocketAddr::new(ip, ent.service_port));
+		let ip = ent.address.parse::<IpAddr>().ok();
+		let pubkey = ent
+			.node_meta
+			.get("pubkey")
+			.map(|k| hex::decode(&k).ok())
+			.flatten()
+			.map(|k| NodeID::from_slice(&k[..]))
+			.flatten();
+		if let (Some(ip), Some(pubkey)) = (ip, pubkey) {
+			ret.push((pubkey, SocketAddr::new(ip, ent.service_port)));
+		} else {
+			warn!(
+				"Could not process node spec from Consul: {:?} (invalid IP or public key)",
+				ent
+			);
+		}
 	}
 	debug!("Got nodes from Consul: {:?}", ret);
 
 	Ok(ret)
 }
+
+// ---- PUBLISHING TO CONSUL CATALOG ----
+
+#[derive(Serialize, Clone, Debug)]
+struct ConsulPublishEntry {
+	#[serde(rename = "Node")]
+	node: String,
+	#[serde(rename = "Address")]
+	address: IpAddr,
+	#[serde(rename = "NodeMeta")]
+	node_meta: HashMap<String, String>,
+	#[serde(rename = "Service")]
+	service: ConsulPublishService,
+}
+
+#[derive(Serialize, Clone, Debug)]
+struct ConsulPublishService {
+	#[serde(rename = "ID")]
+	service_id: String,
+	#[serde(rename = "Service")]
+	service_name: String,
+	#[serde(rename = "Tags")]
+	tags: Vec<String>,
+	#[serde(rename = "Address")]
+	address: IpAddr,
+	#[serde(rename = "Port")]
+	port: u16,
+}
+
+pub async fn publish_consul_service(
+	consul_host: &str,
+	consul_service_name: &str,
+	node_id: NodeID,
+	hostname: &str,
+	rpc_public_addr: SocketAddr,
+) -> Result<(), Error> {
+	let node = format!("garage:{}", hex::encode(&node_id[..8]));
+
+	let advertisment = ConsulPublishEntry {
+		node: node.clone(),
+		address: rpc_public_addr.ip(),
+		node_meta: [
+			("pubkey".to_string(), hex::encode(node_id)),
+			("hostname".to_string(), hostname.to_string()),
+		]
+		.iter()
+		.cloned()
+		.collect(),
+		service: ConsulPublishService {
+			service_id: node.clone(),
+			service_name: consul_service_name.to_string(),
+			tags: vec!["advertised-by-garage".into(), hostname.into()],
+			address: rpc_public_addr.ip(),
+			port: rpc_public_addr.port(),
+		},
+	};
+
+	let url = format!("http://{}/v1/catalog/register", consul_host);
+	let req_body = serde_json::to_string(&advertisment)?;
+	debug!("Request body for consul adv: {}", req_body);
+
+	let req = Request::builder()
+		.uri(url)
+		.method(Method::PUT)
+		.body(Body::from(req_body))?;
+
+	let client = Client::new();
+
+	let resp = client.request(req).await?;
+	debug!("Response of advertising to Consul: {:?}", resp);
+	let resp_code = resp.status();
+	debug!(
+		"{}",
+		std::str::from_utf8(&hyper::body::to_bytes(resp.into_body()).await?)
+			.unwrap_or("<invalid utf8>")
+	);
+
+	if resp_code != StatusCode::OK {
+		return Err(Error::Message(format!("HTTP error {}", resp_code)));
+	}
+
+	Ok(())
+}

src/rpc/lib.rs

@@ -4,10 +4,10 @@
 extern crate log;
 
 mod consul;
-pub(crate) mod tls_util;
 
-pub mod membership;
 pub mod ring;
+pub mod system;
 
-pub mod rpc_client;
-pub mod rpc_server;
+pub mod rpc_helper;
+
+pub use rpc_helper::*;

src/rpc/membership.rs

@@ -1,706 +0,0 @@
-//! Module containing structs related to membership management
-use std::collections::HashMap;
-use std::fmt::Write as FmtWrite;
-use std::io::{Read, Write};
-use std::net::{IpAddr, SocketAddr};
-use std::path::{Path, PathBuf};
-use std::sync::atomic::{AtomicUsize, Ordering};
-use std::sync::Arc;
-use std::time::Duration;
-
-use futures::future::join_all;
-use futures::select;
-use futures_util::future::*;
-use serde::{Deserialize, Serialize};
-use tokio::sync::watch;
-use tokio::sync::Mutex;
-
-use garage_util::background::BackgroundRunner;
-use garage_util::data::*;
-use garage_util::error::Error;
-use garage_util::persister::Persister;
-use garage_util::time::*;
-
-use crate::consul::get_consul_nodes;
-use crate::ring::*;
-use crate::rpc_client::*;
-use crate::rpc_server::*;
-
-const PING_INTERVAL: Duration = Duration::from_secs(10);
-const DISCOVERY_INTERVAL: Duration = Duration::from_secs(60);
-const PING_TIMEOUT: Duration = Duration::from_secs(2);
-const MAX_FAILURES_BEFORE_CONSIDERED_DOWN: usize = 5;
-
-/// RPC endpoint used for calls related to membership
-pub const MEMBERSHIP_RPC_PATH: &str = "_membership";
-
-/// RPC messages related to membership
-#[derive(Debug, Serialize, Deserialize)]
-pub enum Message {
-	/// Response to successfull advertisements
-	Ok,
-	/// Message sent to detect other nodes status
-	Ping(PingMessage),
-	/// Ask other node for the nodes it knows. Answered with AdvertiseNodesUp
-	PullStatus,
-	/// Ask other node its config. Answered with AdvertiseConfig
-	PullConfig,
-	/// Advertisement of nodes the host knows up. Sent spontanously or in response to PullStatus
-	AdvertiseNodesUp(Vec<AdvertisedNode>),
-	/// Advertisement of nodes config. Sent spontanously or in response to PullConfig
-	AdvertiseConfig(NetworkConfig),
-}
-
-impl RpcMessage for Message {}
-
-/// A ping, containing informations about status and config
-#[derive(Debug, Serialize, Deserialize)]
-pub struct PingMessage {
-	id: Uuid,
-	rpc_port: u16,
-
-	status_hash: Hash,
-	config_version: u64,
-
-	state_info: StateInfo,
-}
-
-/// A node advertisement
-#[derive(Clone, Debug, Serialize, Deserialize)]
-pub struct AdvertisedNode {
-	/// Id of the node this advertisement relates to
-	pub id: Uuid,
-	/// IP and port of the node
-	pub addr: SocketAddr,
-
-	/// Is the node considered up
-	pub is_up: bool,
-	/// When was the node last seen up, in milliseconds since UNIX epoch
-	pub last_seen: u64,
-
-	pub state_info: StateInfo,
-}
-
-/// This node's membership manager
-pub struct System {
-	/// The id of this node
-	pub id: Uuid,
-
-	persist_config: Persister<NetworkConfig>,
-	persist_status: Persister<Vec<AdvertisedNode>>,
-	rpc_local_port: u16,
-
-	state_info: StateInfo,
-
-	rpc_http_client: Arc<RpcHttpClient>,
-	rpc_client: Arc<RpcClient<Message>>,
-
-	replication_factor: usize,
-	pub(crate) status: watch::Receiver<Arc<Status>>,
-	/// The ring
-	pub ring: watch::Receiver<Arc<Ring>>,
-
-	update_lock: Mutex<Updaters>,
-
-	/// The job runner of this node
-	pub background: Arc<BackgroundRunner>,
-}
-
-struct Updaters {
-	update_status: watch::Sender<Arc<Status>>,
-	update_ring: watch::Sender<Arc<Ring>>,
-}
-
-/// The status of each nodes, viewed by this node
-#[derive(Debug, Clone)]
-pub struct Status {
-	/// Mapping of each node id to its known status
-	pub nodes: HashMap<Uuid, Arc<StatusEntry>>,
-	/// Hash of `nodes`, used to detect when nodes have different views of the cluster
-	pub hash: Hash,
-}
-
-/// The status of a single node
-#[derive(Debug)]
-pub struct StatusEntry {
-	/// The IP and port used to connect to this node
-	pub addr: SocketAddr,
-	/// Last time this node was seen
-	pub last_seen: u64,
-	/// Number of consecutive pings sent without reply to this node
-	pub num_failures: AtomicUsize,
-	pub state_info: StateInfo,
-}
-
-impl StatusEntry {
-	/// is the node associated to this entry considered up
-	pub fn is_up(&self) -> bool {
-		self.num_failures.load(Ordering::SeqCst) < MAX_FAILURES_BEFORE_CONSIDERED_DOWN
-	}
-}
-
-#[derive(Debug, Clone, Serialize, Deserialize)]
-pub struct StateInfo {
-	pub hostname: String,
-}
-
-impl Status {
-	fn handle_ping(&mut self, ip: IpAddr, info: &PingMessage) -> bool {
-		let addr = SocketAddr::new(ip, info.rpc_port);
-		let old_status = self.nodes.insert(
-			info.id,
-			Arc::new(StatusEntry {
-				addr,
-				last_seen: now_msec(),
-				num_failures: AtomicUsize::from(0),
-				state_info: info.state_info.clone(),
-			}),
-		);
-		match old_status {
-			None => {
-				info!("Newly pingable node: {}", hex::encode(&info.id));
-				true
-			}
-			Some(x) => x.addr != addr,
-		}
-	}
-
-	fn recalculate_hash(&mut self) {
-		let mut nodes = self.nodes.iter().collect::<Vec<_>>();
-		nodes.sort_unstable_by_key(|(id, _status)| *id);
-
-		let mut nodes_txt = String::new();
-		debug!("Current set of pingable nodes: --");
-		for (id, status) in nodes {
-			debug!("{} {}", hex::encode(&id), status.addr);
-			writeln!(&mut nodes_txt, "{} {}", hex::encode(&id), status.addr).unwrap();
-		}
-		debug!("END --");
-		self.hash = blake2sum(nodes_txt.as_bytes());
-	}
-
-	fn to_serializable_membership(&self, system: &System) -> Vec<AdvertisedNode> {
-		let mut mem = vec![];
-		for (node, status) in self.nodes.iter() {
-			let state_info = if *node == system.id {
-				system.state_info.clone()
-			} else {
-				status.state_info.clone()
-			};
-			mem.push(AdvertisedNode {
-				id: *node,
-				addr: status.addr,
-				is_up: status.is_up(),
-				last_seen: status.last_seen,
-				state_info,
-			});
-		}
-		mem
-	}
-}
-
-fn gen_node_id(metadata_dir: &Path) -> Result<Uuid, Error> {
-	let mut id_file = metadata_dir.to_path_buf();
-	id_file.push("node_id");
-	if id_file.as_path().exists() {
-		let mut f = std::fs::File::open(id_file.as_path())?;
-		let mut d = vec![];
-		f.read_to_end(&mut d)?;
-		if d.len() != 32 {
-			return Err(Error::Message("Corrupt node_id file".to_string()));
-		}
-
-		let mut id = [0u8; 32];
-		id.copy_from_slice(&d[..]);
-		Ok(id.into())
-	} else {
-		let id = gen_uuid();
-
-		let mut f = std::fs::File::create(id_file.as_path())?;
-		f.write_all(id.as_slice())?;
-		Ok(id)
-	}
-}
-
-impl System {
-	/// Create this node's membership manager
-	pub fn new(
-		metadata_dir: PathBuf,
-		rpc_http_client: Arc<RpcHttpClient>,
-		background: Arc<BackgroundRunner>,
-		rpc_server: &mut RpcServer,
-		replication_factor: usize,
-	) -> Arc<Self> {
-		let id = gen_node_id(&metadata_dir).expect("Unable to read or generate node ID");
-		info!("Node ID: {}", hex::encode(&id));
-
-		let persist_config = Persister::new(&metadata_dir, "network_config");
-		let persist_status = Persister::new(&metadata_dir, "peer_info");
-
-		let net_config = match persist_config.load() {
-			Ok(x) => x,
-			Err(e) => {
-				match Persister::<garage_rpc_021::ring::NetworkConfig>::new(
-					&metadata_dir,
-					"network_config",
-				)
-				.load()
-				{
-					Ok(old_config) => NetworkConfig::migrate_from_021(old_config),
-					Err(e2) => {
-						info!(
-							"No valid previous network configuration stored ({}, {}), starting fresh.",
-							e, e2
-						);
-						NetworkConfig::new()
-					}
-				}
-			}
-		};
-
-		let mut status = Status {
-			nodes: HashMap::new(),
-			hash: Hash::default(),
-		};
-		status.recalculate_hash();
-		let (update_status, status) = watch::channel(Arc::new(status));
-
-		let state_info = StateInfo {
-			hostname: gethostname::gethostname()
-				.into_string()
-				.unwrap_or_else(|_| "<invalid utf-8>".to_string()),
-		};
-
-		let ring = Ring::new(net_config, replication_factor);
-		let (update_ring, ring) = watch::channel(Arc::new(ring));
-
-		let rpc_path = MEMBERSHIP_RPC_PATH.to_string();
-		let rpc_client = RpcClient::new(
-			RpcAddrClient::<Message>::new(rpc_http_client.clone(), rpc_path.clone()),
-			background.clone(),
-			status.clone(),
-		);
-
-		let sys = Arc::new(System {
-			id,
-			persist_config,
-			persist_status,
-			rpc_local_port: rpc_server.bind_addr.port(),
-			state_info,
-			rpc_http_client,
-			rpc_client,
-			replication_factor,
-			status,
-			ring,
-			update_lock: Mutex::new(Updaters {
-				update_status,
-				update_ring,
-			}),
-			background,
-		});
-		sys.clone().register_handler(rpc_server, rpc_path);
-		sys
-	}
-
-	fn register_handler(self: Arc<Self>, rpc_server: &mut RpcServer, path: String) {
-		rpc_server.add_handler::<Message, _, _>(path, move |msg, addr| {
-			let self2 = self.clone();
-			async move {
-				match msg {
-					Message::Ping(ping) => self2.handle_ping(&addr, &ping).await,
-
-					Message::PullStatus => Ok(self2.handle_pull_status()),
-					Message::PullConfig => Ok(self2.handle_pull_config()),
-					Message::AdvertiseNodesUp(adv) => self2.handle_advertise_nodes_up(&adv).await,
-					Message::AdvertiseConfig(adv) => self2.handle_advertise_config(&adv).await,
-
-					_ => Err(Error::BadRpc("Unexpected RPC message".to_string())),
-				}
-			}
-		});
-	}
-
-	/// Get an RPC client
-	pub fn rpc_client<M: RpcMessage + 'static>(self: &Arc<Self>, path: &str) -> Arc<RpcClient<M>> {
-		RpcClient::new(
-			RpcAddrClient::new(self.rpc_http_client.clone(), path.to_string()),
-			self.background.clone(),
-			self.status.clone(),
-		)
-	}
-
-	/// Save network configuration to disc
-	async fn save_network_config(self: Arc<Self>) -> Result<(), Error> {
-		let ring = self.ring.borrow().clone();
-		self.persist_config
-			.save_async(&ring.config)
-			.await
-			.expect("Cannot save current cluster configuration");
-		Ok(())
-	}
-
-	fn make_ping(&self) -> Message {
-		let status = self.status.borrow().clone();
-		let ring = self.ring.borrow().clone();
-		Message::Ping(PingMessage {
-			id: self.id,
-			rpc_port: self.rpc_local_port,
-			status_hash: status.hash,
-			config_version: ring.config.version,
-			state_info: self.state_info.clone(),
-		})
-	}
-
-	async fn broadcast(self: Arc<Self>, msg: Message, timeout: Duration) {
-		let status = self.status.borrow().clone();
-		let to = status
-			.nodes
-			.keys()
-			.filter(|x| **x != self.id)
-			.cloned()
-			.collect::<Vec<_>>();
-		self.rpc_client.call_many(&to[..], msg, timeout).await;
-	}
-
-	/// Perform bootstraping, starting the ping loop
-	pub async fn bootstrap(
-		self: Arc<Self>,
-		peers: Vec<SocketAddr>,
-		consul_host: Option<String>,
-		consul_service_name: Option<String>,
-	) {
-		let self2 = self.clone();
-		self.background
-			.spawn_worker("discovery loop".to_string(), |stop_signal| {
-				self2.discovery_loop(peers, consul_host, consul_service_name, stop_signal)
-			});
-
-		let self2 = self.clone();
-		self.background
-			.spawn_worker("ping loop".to_string(), |stop_signal| {
-				self2.ping_loop(stop_signal)
-			});
-	}
-
-	async fn ping_nodes(self: Arc<Self>, peers: Vec<(SocketAddr, Option<Uuid>)>) {
-		let ping_msg = self.make_ping();
-		let ping_resps = join_all(peers.iter().map(|(addr, id_option)| {
-			let sys = self.clone();
-			let ping_msg_ref = &ping_msg;
-			async move {
-				(
-					id_option,
-					addr,
-					sys.rpc_client
-						.by_addr()
-						.call(&addr, ping_msg_ref, PING_TIMEOUT)
-						.await,
-				)
-			}
-		}))
-		.await;
-
-		let update_locked = self.update_lock.lock().await;
-		let mut status: Status = self.status.borrow().as_ref().clone();
-		let ring = self.ring.borrow().clone();
-
-		let mut has_changes = false;
-		let mut to_advertise = vec![];
-
-		for (id_option, addr, ping_resp) in ping_resps {
-			if let Ok(Ok(Message::Ping(info))) = ping_resp {
-				let is_new = status.handle_ping(addr.ip(), &info);
-				if is_new {
-					has_changes = true;
-					to_advertise.push(AdvertisedNode {
-						id: info.id,
-						addr: *addr,
-						is_up: true,
-						last_seen: now_msec(),
-						state_info: info.state_info.clone(),
-					});
-				}
-				if is_new || status.hash != info.status_hash {
-					self.background
-						.spawn_cancellable(self.clone().pull_status(info.id).map(Ok));
-				}
-				if is_new || ring.config.version < info.config_version {
-					self.background
-						.spawn_cancellable(self.clone().pull_config(info.id).map(Ok));
-				}
-			} else if let Some(id) = id_option {
-				if let Some(st) = status.nodes.get_mut(id) {
-					// we need to increment failure counter as call was done using by_addr so the
-					// counter was not auto-incremented
-					st.num_failures.fetch_add(1, Ordering::SeqCst);
-					if !st.is_up() {
-						warn!("Node {:?} seems to be down.", id);
-						if !ring.config.members.contains_key(id) {
-							info!("Removing node {:?} from status (not in config and not responding to pings anymore)", id);
-							status.nodes.remove(&id);
-							has_changes = true;
-						}
-					}
-				}
-			}
-		}
-		if has_changes {
-			status.recalculate_hash();
-		}
-		self.update_status(&update_locked, status).await;
-		drop(update_locked);
-
-		if !to_advertise.is_empty() {
-			self.broadcast(Message::AdvertiseNodesUp(to_advertise), PING_TIMEOUT)
-				.await;
-		}
-	}
-
-	async fn handle_ping(
-		self: Arc<Self>,
-		from: &SocketAddr,
-		ping: &PingMessage,
-	) -> Result<Message, Error> {
-		let update_locked = self.update_lock.lock().await;
-		let mut status: Status = self.status.borrow().as_ref().clone();
-
-		let is_new = status.handle_ping(from.ip(), ping);
-		if is_new {
-			status.recalculate_hash();
-		}
-		let status_hash = status.hash;
-		let config_version = self.ring.borrow().config.version;
-
-		self.update_status(&update_locked, status).await;
-		drop(update_locked);
-
-		if is_new || status_hash != ping.status_hash {
-			self.background
-				.spawn_cancellable(self.clone().pull_status(ping.id).map(Ok));
-		}
-		if is_new || config_version < ping.config_version {
-			self.background
-				.spawn_cancellable(self.clone().pull_config(ping.id).map(Ok));
-		}
-
-		Ok(self.make_ping())
-	}
-
-	fn handle_pull_status(&self) -> Message {
-		Message::AdvertiseNodesUp(self.status.borrow().to_serializable_membership(self))
-	}
-
-	fn handle_pull_config(&self) -> Message {
-		let ring = self.ring.borrow().clone();
-		Message::AdvertiseConfig(ring.config.clone())
-	}
-
-	async fn handle_advertise_nodes_up(
-		self: Arc<Self>,
-		adv: &[AdvertisedNode],
-	) -> Result<Message, Error> {
-		let mut to_ping = vec![];
-
-		let update_lock = self.update_lock.lock().await;
-		let mut status: Status = self.status.borrow().as_ref().clone();
-		let mut has_changed = false;
-
-		for node in adv.iter() {
-			if node.id == self.id {
-				// learn our own ip address
-				let self_addr = SocketAddr::new(node.addr.ip(), self.rpc_local_port);
-				let old_self = status.nodes.insert(
-					node.id,
-					Arc::new(StatusEntry {
-						addr: self_addr,
-						last_seen: now_msec(),
-						num_failures: AtomicUsize::from(0),
-						state_info: self.state_info.clone(),
-					}),
-				);
-				has_changed = match old_self {
-					None => true,
-					Some(x) => x.addr != self_addr,
-				};
-			} else {
-				let ping_them = match status.nodes.get(&node.id) {
-					// Case 1: new node
-					None => true,
-					// Case 2: the node might have changed address
-					Some(our_node) => node.is_up && !our_node.is_up() && our_node.addr != node.addr,
-				};
-				if ping_them {
-					to_ping.push((node.addr, Some(node.id)));
-				}
-			}
-		}
-		if has_changed {
-			status.recalculate_hash();
-		}
-		self.update_status(&update_lock, status).await;
-		drop(update_lock);
-
-		if !to_ping.is_empty() {
-			self.background
-				.spawn_cancellable(self.clone().ping_nodes(to_ping).map(Ok));
-		}
-
-		Ok(Message::Ok)
-	}
-
-	async fn handle_advertise_config(
-		self: Arc<Self>,
-		adv: &NetworkConfig,
-	) -> Result<Message, Error> {
-		let update_lock = self.update_lock.lock().await;
-		let ring: Arc<Ring> = self.ring.borrow().clone();
-
-		if adv.version > ring.config.version {
-			let ring = Ring::new(adv.clone(), self.replication_factor);
-			update_lock.update_ring.send(Arc::new(ring))?;
-			drop(update_lock);
-
-			self.background.spawn_cancellable(
-				self.clone()
-					.broadcast(Message::AdvertiseConfig(adv.clone()), PING_TIMEOUT)
-					.map(Ok),
-			);
-			self.background.spawn(self.clone().save_network_config());
-		}
-
-		Ok(Message::Ok)
-	}
-
-	async fn ping_loop(self: Arc<Self>, mut stop_signal: watch::Receiver<bool>) {
-		while !*stop_signal.borrow() {
-			let restart_at = tokio::time::sleep(PING_INTERVAL);
-
-			let status = self.status.borrow().clone();
-			let ping_addrs = status
-				.nodes
-				.iter()
-				.filter(|(id, _)| **id != self.id)
-				.map(|(id, status)| (status.addr, Some(*id)))
-				.collect::<Vec<_>>();
-
-			self.clone().ping_nodes(ping_addrs).await;
-
-			select! {
-				_ = restart_at.fuse() => {},
-				_ = stop_signal.changed().fuse() => {},
-			}
-		}
-	}
-
-	async fn discovery_loop(
-		self: Arc<Self>,
-		bootstrap_peers: Vec<SocketAddr>,
-		consul_host: Option<String>,
-		consul_service_name: Option<String>,
-		mut stop_signal: watch::Receiver<bool>,
-	) {
-		let consul_config = match (consul_host, consul_service_name) {
-			(Some(ch), Some(csn)) => Some((ch, csn)),
-			_ => None,
-		};
-
-		while !*stop_signal.borrow() {
-			let not_configured = self.ring.borrow().config.members.is_empty();
-			let no_peers = self.status.borrow().nodes.len() < 3;
-			let bad_peers = self
-				.status
-				.borrow()
-				.nodes
-				.iter()
-				.filter(|(_, v)| v.is_up())
-				.count() != self.ring.borrow().config.members.len();
-
-			if not_configured || no_peers || bad_peers {
-				info!("Doing a bootstrap/discovery step (not_configured: {}, no_peers: {}, bad_peers: {})", not_configured, no_peers, bad_peers);
-
-				let mut ping_list = bootstrap_peers
-					.iter()
-					.map(|ip| (*ip, None))
-					.collect::<Vec<_>>();
-
-				if let Ok(peers) = self.persist_status.load_async().await {
-					ping_list.extend(peers.iter().map(|x| (x.addr, Some(x.id))));
-				}
-
-				if let Some((consul_host, consul_service_name)) = &consul_config {
-					match get_consul_nodes(consul_host, consul_service_name).await {
-						Ok(node_list) => {
-							ping_list.extend(node_list.iter().map(|a| (*a, None)));
-						}
-						Err(e) => {
-							warn!("Could not retrieve node list from Consul: {}", e);
-						}
-					}
-				}
-
-				self.clone().ping_nodes(ping_list).await;
-			}
-
-			let restart_at = tokio::time::sleep(DISCOVERY_INTERVAL);
-			select! {
-				_ = restart_at.fuse() => {},
-				_ = stop_signal.changed().fuse() => {},
-			}
-		}
-	}
-
-	// for some reason fixing this is causing compilation error, see https://github.com/rust-lang/rust-clippy/issues/7052
-	#[allow(clippy::manual_async_fn)]
-	fn pull_status(
-		self: Arc<Self>,
-		peer: Uuid,
-	) -> impl futures::future::Future<Output = ()> + Send + 'static {
-		async move {
-			let resp = self
-				.rpc_client
-				.call(peer, Message::PullStatus, PING_TIMEOUT)
-				.await;
-			if let Ok(Message::AdvertiseNodesUp(nodes)) = resp {
-				let _: Result<_, _> = self.handle_advertise_nodes_up(&nodes).await;
-			}
-		}
-	}
-
-	async fn pull_config(self: Arc<Self>, peer: Uuid) {
-		let resp = self
-			.rpc_client
-			.call(peer, Message::PullConfig, PING_TIMEOUT)
-			.await;
-		if let Ok(Message::AdvertiseConfig(config)) = resp {
-			let _: Result<_, _> = self.handle_advertise_config(&config).await;
-		}
-	}
-
-	async fn update_status(self: &Arc<Self>, updaters: &Updaters, status: Status) {
-		if status.hash != self.status.borrow().hash {
-			let mut list = status.to_serializable_membership(&self);
-
-			// Combine with old peer list to make sure no peer is lost
-			if let Ok(old_list) = self.persist_status.load_async().await {
-				for pp in old_list {
-					if !list.iter().any(|np| pp.id == np.id) {
-						list.push(pp);
-					}
-				}
-			}
-
-			if !list.is_empty() {
-				info!("Persisting new peer list ({} peers)", list.len());
-				self.persist_status
-					.save_async(&list)
-					.await
-					.expect("Unable to persist peer list");
-			}
-		}
-
-		updaters
-			.update_status
-			.send(Arc::new(status))
-			.expect("Could not update internal membership status");
-	}
-}

src/rpc/ring.rs

@@ -38,31 +38,6 @@ impl NetworkConfig {
 			version: 0,
 		}
 	}
-
-	pub(crate) fn migrate_from_021(old: garage_rpc_021::ring::NetworkConfig) -> Self {
-		let members = old
-			.members
-			.into_iter()
-			.map(|(id, conf)| {
-				(
-					Hash::try_from(id.as_slice()).unwrap(),
-					NetworkConfigEntry {
-						zone: conf.datacenter,
-						capacity: if conf.capacity == 0 {
-							None
-						} else {
-							Some(conf.capacity)
-						},
-						tag: conf.tag,
-					},
-				)
-			})
-			.collect();
-		Self {
-			members,
-			version: old.version,
-		}
-	}
 }
 
 /// The overall configuration of one (possibly remote) node

src/rpc/rpc_client.rs

@@ -1,365 +0,0 @@
-//! Contain structs related to making RPCs
-use std::borrow::Borrow;
-use std::marker::PhantomData;
-use std::net::SocketAddr;
-use std::pin::Pin;
-use std::sync::atomic::Ordering;
-use std::sync::Arc;
-use std::time::Duration;
-
-use arc_swap::ArcSwapOption;
-use futures::future::Future;
-use futures::stream::futures_unordered::FuturesUnordered;
-use futures::stream::StreamExt;
-use futures_util::future::FutureExt;
-use hyper::client::{Client, HttpConnector};
-use hyper::{Body, Method, Request};
-use tokio::sync::{watch, Semaphore};
-
-use garage_util::background::BackgroundRunner;
-use garage_util::config::TlsConfig;
-use garage_util::data::*;
-use garage_util::error::{Error, RpcError};
-
-use crate::membership::Status;
-use crate::rpc_server::RpcMessage;
-use crate::tls_util;
-
-const DEFAULT_TIMEOUT: Duration = Duration::from_secs(10);
-
-/// Strategy to apply when making RPC
-#[derive(Copy, Clone)]
-pub struct RequestStrategy {
-	/// Max time to wait for reponse
-	pub rs_timeout: Duration,
-	/// Min number of response to consider the request successful
-	pub rs_quorum: usize,
-	/// Should requests be dropped after enough response are received
-	pub rs_interrupt_after_quorum: bool,
-}
-
-impl RequestStrategy {
-	/// Create a RequestStrategy with default timeout and not interrupting when quorum reached
-	pub fn with_quorum(quorum: usize) -> Self {
-		RequestStrategy {
-			rs_timeout: DEFAULT_TIMEOUT,
-			rs_quorum: quorum,
-			rs_interrupt_after_quorum: false,
-		}
-	}
-	/// Set timeout of the strategy
-	pub fn with_timeout(mut self, timeout: Duration) -> Self {
-		self.rs_timeout = timeout;
-		self
-	}
-	/// Set if requests can be dropped after quorum has been reached
-	/// In general true for read requests, and false for write
-	pub fn interrupt_after_quorum(mut self, interrupt: bool) -> Self {
-		self.rs_interrupt_after_quorum = interrupt;
-		self
-	}
-}
-
-/// Shortcut for a boxed async function taking a message, and resolving to another message or an
-/// error
-pub type LocalHandlerFn<M> =
-	Box<dyn Fn(Arc<M>) -> Pin<Box<dyn Future<Output = Result<M, Error>> + Send>> + Send + Sync>;
-
-/// Client used to send RPC
-pub struct RpcClient<M: RpcMessage> {
-	status: watch::Receiver<Arc<Status>>,
-	background: Arc<BackgroundRunner>,
-
-	local_handler: ArcSwapOption<(Uuid, LocalHandlerFn<M>)>,
-
-	rpc_addr_client: RpcAddrClient<M>,
-}
-
-impl<M: RpcMessage + 'static> RpcClient<M> {
-	/// Create a new RpcClient from an address, a job runner, and the status of all RPC servers
-	pub fn new(
-		rac: RpcAddrClient<M>,
-		background: Arc<BackgroundRunner>,
-		status: watch::Receiver<Arc<Status>>,
-	) -> Arc<Self> {
-		Arc::new(Self {
-			rpc_addr_client: rac,
-			background,
-			status,
-			local_handler: ArcSwapOption::new(None),
-		})
-	}
-
-	/// Set the local handler, to process RPC to this node without network usage
-	pub fn set_local_handler<F, Fut>(&self, my_id: Uuid, handler: F)
-	where
-		F: Fn(Arc<M>) -> Fut + Send + Sync + 'static,
-		Fut: Future<Output = Result<M, Error>> + Send + 'static,
-	{
-		let handler_arc = Arc::new(handler);
-		let handler: LocalHandlerFn<M> = Box::new(move |msg| {
-			let handler_arc2 = handler_arc.clone();
-			Box::pin(async move { handler_arc2(msg).await })
-		});
-		self.local_handler.swap(Some(Arc::new((my_id, handler))));
-	}
-
-	/// Get a RPC client to make calls using node's SocketAddr instead of its ID
-	pub fn by_addr(&self) -> &RpcAddrClient<M> {
-		&self.rpc_addr_client
-	}
-
-	/// Make a RPC call
-	pub async fn call(&self, to: Uuid, msg: M, timeout: Duration) -> Result<M, Error> {
-		self.call_arc(to, Arc::new(msg), timeout).await
-	}
-
-	/// Make a RPC call from a message stored in an Arc
-	pub async fn call_arc(&self, to: Uuid, msg: Arc<M>, timeout: Duration) -> Result<M, Error> {
-		if let Some(lh) = self.local_handler.load_full() {
-			let (my_id, local_handler) = lh.as_ref();
-			if to.borrow() == my_id {
-				return local_handler(msg).await;
-			}
-		}
-		let status = self.status.borrow().clone();
-		let node_status = match status.nodes.get(&to) {
-			Some(node_status) => {
-				if node_status.is_up() {
-					node_status
-				} else {
-					return Err(Error::from(RpcError::NodeDown(to)));
-				}
-			}
-			None => {
-				return Err(Error::Message(format!(
-					"Peer ID not found: {:?}",
-					to.borrow()
-				)))
-			}
-		};
-		match self
-			.rpc_addr_client
-			.call(&node_status.addr, msg, timeout)
-			.await
-		{
-			Err(rpc_error) => {
-				node_status.num_failures.fetch_add(1, Ordering::SeqCst);
-				Err(Error::from(rpc_error))
-			}
-			Ok(x) => x,
-		}
-	}
-
-	/// Make a RPC call to multiple servers, returning a Vec containing each result
-	pub async fn call_many(&self, to: &[Uuid], msg: M, timeout: Duration) -> Vec<Result<M, Error>> {
-		let msg = Arc::new(msg);
-		let mut resp_stream = to
-			.iter()
-			.map(|to| self.call_arc(*to, msg.clone(), timeout))
-			.collect::<FuturesUnordered<_>>();
-
-		let mut results = vec![];
-		while let Some(resp) = resp_stream.next().await {
-			results.push(resp);
-		}
-		results
-	}
-
-	/// Make a RPC call to multiple servers, returning either a Vec of responses, or an error if
-	/// strategy could not be respected due to too many errors
-	pub async fn try_call_many(
-		self: &Arc<Self>,
-		to: &[Uuid],
-		msg: M,
-		strategy: RequestStrategy,
-	) -> Result<Vec<M>, Error> {
-		let timeout = strategy.rs_timeout;
-
-		let msg = Arc::new(msg);
-		let mut resp_stream = to
-			.to_vec()
-			.into_iter()
-			.map(|to| {
-				let self2 = self.clone();
-				let msg = msg.clone();
-				async move { self2.call_arc(to, msg, timeout).await }
-			})
-			.collect::<FuturesUnordered<_>>();
-
-		let mut results = vec![];
-		let mut errors = vec![];
-
-		while let Some(resp) = resp_stream.next().await {
-			match resp {
-				Ok(msg) => {
-					results.push(msg);
-					if results.len() >= strategy.rs_quorum {
-						break;
-					}
-				}
-				Err(e) => {
-					errors.push(e);
-				}
-			}
-		}
-
-		if results.len() >= strategy.rs_quorum {
-			// Continue requests in background.
-			// Continue the remaining requests immediately using tokio::spawn
-			// but enqueue a task in the background runner
-			// to ensure that the process won't exit until the requests are done
-			// (if we had just enqueued the resp_stream.collect directly in the background runner,
-			// the requests might have been put on hold in the background runner's queue,
-			// in which case they might timeout or otherwise fail)
-			if !strategy.rs_interrupt_after_quorum {
-				let wait_finished_fut = tokio::spawn(async move {
-					resp_stream.collect::<Vec<_>>().await;
-				});
-				self.background.spawn(wait_finished_fut.map(|_| Ok(())));
-			}
-
-			Ok(results)
-		} else {
-			let errors = errors.iter().map(|e| format!("{}", e)).collect::<Vec<_>>();
-			Err(Error::from(RpcError::TooManyErrors(errors)))
-		}
-	}
-}
-
-/// Thin wrapper arround an `RpcHttpClient` specifying the path of the request
-pub struct RpcAddrClient<M: RpcMessage> {
-	phantom: PhantomData<M>,
-
-	http_client: Arc<RpcHttpClient>,
-	path: String,
-}
-
-impl<M: RpcMessage> RpcAddrClient<M> {
-	/// Create an RpcAddrClient from an HTTP client and the endpoint to reach for RPCs
-	pub fn new(http_client: Arc<RpcHttpClient>, path: String) -> Self {
-		Self {
-			phantom: PhantomData::default(),
-			http_client,
-			path,
-		}
-	}
-
-	/// Make a RPC
-	pub async fn call<MB>(
-		&self,
-		to_addr: &SocketAddr,
-		msg: MB,
-		timeout: Duration,
-	) -> Result<Result<M, Error>, RpcError>
-	where
-		MB: Borrow<M>,
-	{
-		self.http_client
-			.call(&self.path, to_addr, msg, timeout)
-			.await
-	}
-}
-
-/// HTTP client used to make RPCs
-pub struct RpcHttpClient {
-	request_limiter: Semaphore,
-	method: ClientMethod,
-}
-
-enum ClientMethod {
-	Http(Client<HttpConnector, hyper::Body>),
-	Https(Client<tls_util::HttpsConnectorFixedDnsname<HttpConnector>, hyper::Body>),
-}
-
-impl RpcHttpClient {
-	/// Create a new RpcHttpClient
-	pub fn new(
-		max_concurrent_requests: usize,
-		tls_config: &Option<TlsConfig>,
-	) -> Result<Self, Error> {
-		let method = if let Some(cf) = tls_config {
-			let ca_certs = tls_util::load_certs(&cf.ca_cert)?;
-			let node_certs = tls_util::load_certs(&cf.node_cert)?;
-			let node_key = tls_util::load_private_key(&cf.node_key)?;
-
-			let mut config = rustls::ClientConfig::new();
-
-			for crt in ca_certs.iter() {
-				config.root_store.add(crt)?;
-			}
-
-			config.set_single_client_cert([&node_certs[..], &ca_certs[..]].concat(), node_key)?;
-
-			let connector =
-				tls_util::HttpsConnectorFixedDnsname::<HttpConnector>::new(config, "garage");
-
-			ClientMethod::Https(Client::builder().build(connector))
-		} else {
-			ClientMethod::Http(Client::new())
-		};
-		Ok(RpcHttpClient {
-			method,
-			request_limiter: Semaphore::new(max_concurrent_requests),
-		})
-	}
-
-	/// Make a RPC
-	async fn call<M, MB>(
-		&self,
-		path: &str,
-		to_addr: &SocketAddr,
-		msg: MB,
-		timeout: Duration,
-	) -> Result<Result<M, Error>, RpcError>
-	where
-		MB: Borrow<M>,
-		M: RpcMessage,
-	{
-		let uri = match self.method {
-			ClientMethod::Http(_) => format!("http://{}/{}", to_addr, path),
-			ClientMethod::Https(_) => format!("https://{}/{}", to_addr, path),
-		};
-
-		let req = Request::builder()
-			.method(Method::POST)
-			.uri(uri)
-			.body(Body::from(rmp_to_vec_all_named(msg.borrow())?))?;
-
-		let resp_fut = match &self.method {
-			ClientMethod::Http(client) => client.request(req).fuse(),
-			ClientMethod::Https(client) => client.request(req).fuse(),
-		};
-
-		trace!("({}) Acquiring request_limiter slot...", path);
-		let slot = self.request_limiter.acquire().await;
-		trace!("({}) Got slot, doing request to {}...", path, to_addr);
-		let resp = tokio::time::timeout(timeout, resp_fut)
-			.await
-			.map_err(|e| {
-				debug!(
-					"RPC timeout to {}: {}",
-					to_addr,
-					debug_serialize(msg.borrow())
-				);
-				e
-			})?
-			.map_err(|e| {
-				warn!(
-					"RPC HTTP client error when connecting to {}: {}",
-					to_addr, e
-				);
-				e
-			})?;
-
-		let status = resp.status();
-		trace!("({}) Request returned, got status {}", path, status);
-		let body = hyper::body::to_bytes(resp.into_body()).await?;
-		drop(slot);
-
-		match rmp_serde::decode::from_read::<_, Result<M, String>>(&body[..])? {
-			Err(e) => Ok(Err(Error::RemoteError(e, status))),
-			Ok(x) => Ok(Ok(x)),
-		}
-	}
-}

src/rpc/rpc_helper.rs

@@ -0,0 +1,343 @@
+//! Contain structs related to making RPCs
+use std::sync::Arc;
+use std::time::Duration;
+
+use futures::future::join_all;
+use futures::stream::futures_unordered::FuturesUnordered;
+use futures::stream::StreamExt;
+use futures_util::future::FutureExt;
+use tokio::select;
+use tokio::sync::{watch, Semaphore};
+
+pub use netapp::endpoint::{Endpoint, EndpointHandler, Message as Rpc};
+use netapp::peering::fullmesh::FullMeshPeeringStrategy;
+pub use netapp::proto::*;
+pub use netapp::{NetApp, NodeID};
+
+use garage_util::background::BackgroundRunner;
+use garage_util::data::*;
+use garage_util::error::Error;
+
+use crate::ring::Ring;
+
+const DEFAULT_TIMEOUT: Duration = Duration::from_secs(10);
+
+// Try to never have more than 200MB of outgoing requests
+// buffered at the same time. Other requests are queued until
+// space is freed.
+const REQUEST_BUFFER_SIZE: usize = 200 * 1024 * 1024;
+
+/// Strategy to apply when making RPC
+#[derive(Copy, Clone)]
+pub struct RequestStrategy {
+	/// Max time to wait for reponse
+	pub rs_timeout: Duration,
+	/// Min number of response to consider the request successful
+	pub rs_quorum: Option<usize>,
+	/// Should requests be dropped after enough response are received
+	pub rs_interrupt_after_quorum: bool,
+	/// Request priority
+	pub rs_priority: RequestPriority,
+}
+
+impl RequestStrategy {
+	/// Create a RequestStrategy with default timeout and not interrupting when quorum reached
+	pub fn with_priority(prio: RequestPriority) -> Self {
+		RequestStrategy {
+			rs_timeout: DEFAULT_TIMEOUT,
+			rs_quorum: None,
+			rs_interrupt_after_quorum: false,
+			rs_priority: prio,
+		}
+	}
+	/// Set quorum to be reached for request
+	pub fn with_quorum(mut self, quorum: usize) -> Self {
+		self.rs_quorum = Some(quorum);
+		self
+	}
+	/// Set timeout of the strategy
+	pub fn with_timeout(mut self, timeout: Duration) -> Self {
+		self.rs_timeout = timeout;
+		self
+	}
+	/// Set if requests can be dropped after quorum has been reached
+	/// In general true for read requests, and false for write
+	pub fn interrupt_after_quorum(mut self, interrupt: bool) -> Self {
+		self.rs_interrupt_after_quorum = interrupt;
+		self
+	}
+}
+
+#[derive(Clone)]
+pub struct RpcHelper(Arc<RpcHelperInner>);
+
+struct RpcHelperInner {
+	our_node_id: Uuid,
+	fullmesh: Arc<FullMeshPeeringStrategy>,
+	background: Arc<BackgroundRunner>,
+	ring: watch::Receiver<Arc<Ring>>,
+	request_buffer_semaphore: Semaphore,
+}
+
+impl RpcHelper {
+	pub(crate) fn new(
+		our_node_id: Uuid,
+		fullmesh: Arc<FullMeshPeeringStrategy>,
+		background: Arc<BackgroundRunner>,
+		ring: watch::Receiver<Arc<Ring>>,
+	) -> Self {
+		Self(Arc::new(RpcHelperInner {
+			our_node_id,
+			fullmesh,
+			background,
+			ring,
+			request_buffer_semaphore: Semaphore::new(REQUEST_BUFFER_SIZE),
+		}))
+	}
+
+	pub async fn call<M, H, S>(
+		&self,
+		endpoint: &Endpoint<M, H>,
+		to: Uuid,
+		msg: M,
+		strat: RequestStrategy,
+	) -> Result<S, Error>
+	where
+		M: Rpc<Response = Result<S, Error>>,
+		H: EndpointHandler<M>,
+	{
+		self.call_arc(endpoint, to, Arc::new(msg), strat).await
+	}
+
+	pub async fn call_arc<M, H, S>(
+		&self,
+		endpoint: &Endpoint<M, H>,
+		to: Uuid,
+		msg: Arc<M>,
+		strat: RequestStrategy,
+	) -> Result<S, Error>
+	where
+		M: Rpc<Response = Result<S, Error>>,
+		H: EndpointHandler<M>,
+	{
+		let msg_size = rmp_to_vec_all_named(&msg)?.len() as u32;
+		let permit = self
+			.0
+			.request_buffer_semaphore
+			.acquire_many(msg_size)
+			.await?;
+
+		let node_id = to.into();
+		select! {
+			res = endpoint.call(&node_id, &msg, strat.rs_priority) => {
+				drop(permit);
+				Ok(res??)
+			}
+			_ = tokio::time::sleep(strat.rs_timeout) => {
+				drop(permit);
+				Err(Error::Timeout)
+			}
+		}
+	}
+
+	pub async fn call_many<M, H, S>(
+		&self,
+		endpoint: &Endpoint<M, H>,
+		to: &[Uuid],
+		msg: M,
+		strat: RequestStrategy,
+	) -> Vec<(Uuid, Result<S, Error>)>
+	where
+		M: Rpc<Response = Result<S, Error>>,
+		H: EndpointHandler<M>,
+	{
+		let msg = Arc::new(msg);
+		let resps = join_all(
+			to.iter()
+				.map(|to| self.call_arc(endpoint, *to, msg.clone(), strat)),
+		)
+		.await;
+		to.iter()
+			.cloned()
+			.zip(resps.into_iter())
+			.collect::<Vec<_>>()
+	}
+
+	pub async fn broadcast<M, H, S>(
+		&self,
+		endpoint: &Endpoint<M, H>,
+		msg: M,
+		strat: RequestStrategy,
+	) -> Vec<(Uuid, Result<S, Error>)>
+	where
+		M: Rpc<Response = Result<S, Error>>,
+		H: EndpointHandler<M>,
+	{
+		let to = self
+			.0
+			.fullmesh
+			.get_peer_list()
+			.iter()
+			.map(|p| p.id.into())
+			.collect::<Vec<_>>();
+		self.call_many(endpoint, &to[..], msg, strat).await
+	}
+
+	/// Make a RPC call to multiple servers, returning either a Vec of responses,
+	/// or an error if quorum could not be reached due to too many errors
+	pub async fn try_call_many<M, H, S>(
+		&self,
+		endpoint: &Arc<Endpoint<M, H>>,
+		to: &[Uuid],
+		msg: M,
+		strategy: RequestStrategy,
+	) -> Result<Vec<S>, Error>
+	where
+		M: Rpc<Response = Result<S, Error>> + 'static,
+		H: EndpointHandler<M> + 'static,
+		S: Send,
+	{
+		let msg = Arc::new(msg);
+
+		// Build future for each request
+		// They are not started now: they are added below in a FuturesUnordered
+		// object that will take care of polling them (see below)
+		let requests = to.iter().cloned().map(|to| {
+			let self2 = self.clone();
+			let msg = msg.clone();
+			let endpoint2 = endpoint.clone();
+			(to, async move {
+				self2.call_arc(&endpoint2, to, msg, strategy).await
+			})
+		});
+		let quorum = strategy.rs_quorum.unwrap_or(to.len());
+
+		// Vectors in which success results and errors will be collected
+		let mut successes = vec![];
+		let mut errors = vec![];
+
+		if strategy.rs_interrupt_after_quorum {
+			// Case 1: once quorum is reached, other requests don't matter.
+			// What we do here is only send the required number of requests
+			// to reach a quorum, priorizing nodes with the lowest latency.
+			// When there are errors, we start new requests to compensate.
+
+			// Retrieve some status variables that we will use to sort requests
+			let peer_list = self.0.fullmesh.get_peer_list();
+			let ring: Arc<Ring> = self.0.ring.borrow().clone();
+			let our_zone = match ring.config.members.get(&self.0.our_node_id) {
+				Some(pc) => &pc.zone,
+				None => "",
+			};
+
+			// Augment requests with some information used to sort them.
+			// The tuples are as follows:
+			//         (is another node?, is another zone?, latency, node ID, request future)
+			// We store all of these tuples in a vec that we can sort.
+			// By sorting this vec, we priorize ourself, then nodes in the same zone,
+			// and within a same zone we priorize nodes with the lowest latency.
+			let mut requests = requests
+				.map(|(to, fut)| {
+					let peer_zone = match ring.config.members.get(&to) {
+						Some(pc) => &pc.zone,
+						None => "",
+					};
+					let peer_avg_ping = peer_list
+						.iter()
+						.find(|x| x.id.as_ref() == to.as_slice())
+						.map(|pi| pi.avg_ping)
+						.flatten()
+						.unwrap_or_else(|| Duration::from_secs(1));
+					(
+						to != self.0.our_node_id,
+						peer_zone != our_zone,
+						peer_avg_ping,
+						to,
+						fut,
+					)
+				})
+				.collect::<Vec<_>>();
+
+			// Sort requests by (priorize ourself, priorize same zone, priorize low latency)
+			requests
+				.sort_by_key(|(diffnode, diffzone, ping, _to, _fut)| (*diffnode, *diffzone, *ping));
+
+			// Make an iterator to take requests in their sorted order
+			let mut requests = requests.into_iter();
+
+			// resp_stream will contain all of the requests that are currently in flight.
+			// (for the moment none, they will be added in the loop below)
+			let mut resp_stream = FuturesUnordered::new();
+
+			// Do some requests and collect results
+			'request_loop: while successes.len() < quorum {
+				// If the current set of requests that are running is not enough to possibly
+				// reach quorum, start some new requests.
+				while successes.len() + resp_stream.len() < quorum {
+					if let Some((_, _, _, _to, fut)) = requests.next() {
+						resp_stream.push(fut);
+					} else {
+						// If we have no request to add, we know that we won't ever
+						// reach quorum: bail out now.
+						break 'request_loop;
+					}
+				}
+				assert!(!resp_stream.is_empty()); // because of loop invariants
+
+				// Wait for one request to terminate
+				match resp_stream.next().await.unwrap() {
+					Ok(msg) => {
+						successes.push(msg);
+					}
+					Err(e) => {
+						errors.push(e);
+					}
+				}
+			}
+		} else {
+			// Case 2: all of the requests need to be sent in all cases,
+			// and need to terminate. (this is the case for writes that
+			// must be spread to n nodes)
+			// Just start all the requests in parallel and return as soon
+			// as the quorum is reached.
+			let mut resp_stream = requests
+				.map(|(_, fut)| fut)
+				.collect::<FuturesUnordered<_>>();
+
+			while let Some(resp) = resp_stream.next().await {
+				match resp {
+					Ok(msg) => {
+						successes.push(msg);
+						if successes.len() >= quorum {
+							break;
+						}
+					}
+					Err(e) => {
+						errors.push(e);
+					}
+				}
+			}
+
+			if !resp_stream.is_empty() {
+				// Continue remaining requests in background.
+				// Continue the remaining requests immediately using tokio::spawn
+				// but enqueue a task in the background runner
+				// to ensure that the process won't exit until the requests are done
+				// (if we had just enqueued the resp_stream.collect directly in the background runner,
+				// the requests might have been put on hold in the background runner's queue,
+				// in which case they might timeout or otherwise fail)
+				let wait_finished_fut = tokio::spawn(async move {
+					resp_stream.collect::<Vec<Result<_, _>>>().await;
+				});
+				self.0.background.spawn(wait_finished_fut.map(|_| Ok(())));
+			}
+		}
+
+		if successes.len() >= quorum {
+			Ok(successes)
+		} else {
+			let errors = errors.iter().map(|e| format!("{}", e)).collect::<Vec<_>>();
+			Err(Error::Quorum(quorum, successes.len(), to.len(), errors))
+		}
+	}
+}

src/rpc/rpc_server.rs

@@ -1,247 +0,0 @@
-//! Contains structs related to receiving RPCs
-use std::collections::HashMap;
-use std::net::SocketAddr;
-use std::pin::Pin;
-use std::sync::Arc;
-use std::time::Instant;
-
-use futures::future::Future;
-use futures_util::future::*;
-use futures_util::stream::*;
-use hyper::server::conn::AddrStream;
-use hyper::service::{make_service_fn, service_fn};
-use hyper::{Body, Method, Request, Response, Server, StatusCode};
-use serde::{Deserialize, Serialize};
-use tokio::net::{TcpListener, TcpStream};
-use tokio_rustls::server::TlsStream;
-use tokio_rustls::TlsAcceptor;
-use tokio_stream::wrappers::TcpListenerStream;
-
-use garage_util::config::TlsConfig;
-use garage_util::data::*;
-use garage_util::error::Error;
-
-use crate::tls_util;
-
-/// Trait for messages that can be sent as RPC
-pub trait RpcMessage: Serialize + for<'de> Deserialize<'de> + Send + Sync {}
-
-type ResponseFuture = Pin<Box<dyn Future<Output = Result<Response<Body>, Error>> + Send>>;
-type Handler = Box<dyn Fn(Request<Body>, SocketAddr) -> ResponseFuture + Send + Sync>;
-
-/// Structure handling RPCs
-pub struct RpcServer {
-	/// The address the RpcServer will bind
-	pub bind_addr: SocketAddr,
-	/// The tls configuration used for RPC
-	pub tls_config: Option<TlsConfig>,
-
-	handlers: HashMap<String, Handler>,
-}
-
-async fn handle_func<M, F, Fut>(
-	handler: Arc<F>,
-	req: Request<Body>,
-	sockaddr: SocketAddr,
-	name: Arc<String>,
-) -> Result<Response<Body>, Error>
-where
-	M: RpcMessage + 'static,
-	F: Fn(M, SocketAddr) -> Fut + Send + Sync + 'static,
-	Fut: Future<Output = Result<M, Error>> + Send + 'static,
-{
-	let begin_time = Instant::now();
-	let whole_body = hyper::body::to_bytes(req.into_body()).await?;
-	let msg = rmp_serde::decode::from_read::<_, M>(&whole_body[..])?;
-
-	trace!(
-		"Request message: {}",
-		serde_json::to_string(&msg)
-			.unwrap_or_else(|_| "<json error>".into())
-			.chars()
-			.take(100)
-			.collect::<String>()
-	);
-
-	match handler(msg, sockaddr).await {
-		Ok(resp) => {
-			let resp_bytes = rmp_to_vec_all_named::<Result<M, String>>(&Ok(resp))?;
-			let rpc_duration = (Instant::now() - begin_time).as_millis();
-			if rpc_duration > 100 {
-				debug!("RPC {} ok, took long: {} ms", name, rpc_duration,);
-			}
-			Ok(Response::new(Body::from(resp_bytes)))
-		}
-		Err(e) => {
-			let err_str = format!("{}", e);
-			let rep_bytes = rmp_to_vec_all_named::<Result<M, String>>(&Err(err_str))?;
-			let mut err_response = Response::new(Body::from(rep_bytes));
-			*err_response.status_mut() = match e {
-				Error::BadRpc(_) => StatusCode::BAD_REQUEST,
-				_ => StatusCode::INTERNAL_SERVER_ERROR,
-			};
-			warn!(
-				"RPC error ({}): {} ({} ms)",
-				name,
-				e,
-				(Instant::now() - begin_time).as_millis(),
-			);
-			Ok(err_response)
-		}
-	}
-}
-
-impl RpcServer {
-	/// Create a new RpcServer
-	pub fn new(bind_addr: SocketAddr, tls_config: Option<TlsConfig>) -> Self {
-		Self {
-			bind_addr,
-			tls_config,
-			handlers: HashMap::new(),
-		}
-	}
-
-	/// Add handler handling request made to `name`
-	pub fn add_handler<M, F, Fut>(&mut self, name: String, handler: F)
-	where
-		M: RpcMessage + 'static,
-		F: Fn(M, SocketAddr) -> Fut + Send + Sync + 'static,
-		Fut: Future<Output = Result<M, Error>> + Send + 'static,
-	{
-		let name2 = Arc::new(name.clone());
-		let handler_arc = Arc::new(handler);
-		let handler = Box::new(move |req: Request<Body>, sockaddr: SocketAddr| {
-			let handler2 = handler_arc.clone();
-			let b: ResponseFuture = Box::pin(handle_func(handler2, req, sockaddr, name2.clone()));
-			b
-		});
-		self.handlers.insert(name, handler);
-	}
-
-	async fn handler(
-		self: Arc<Self>,
-		req: Request<Body>,
-		addr: SocketAddr,
-	) -> Result<Response<Body>, Error> {
-		if req.method() != Method::POST {
-			let mut bad_request = Response::default();
-			*bad_request.status_mut() = StatusCode::BAD_REQUEST;
-			return Ok(bad_request);
-		}
-
-		let path = &req.uri().path()[1..].to_string();
-
-		let handler = match self.handlers.get(path) {
-			Some(h) => h,
-			None => {
-				let mut not_found = Response::default();
-				*not_found.status_mut() = StatusCode::NOT_FOUND;
-				return Ok(not_found);
-			}
-		};
-
-		trace!("({}) Handling request", path);
-
-		let resp_waiter = tokio::spawn(handler(req, addr));
-		match resp_waiter.await {
-			Err(err) => {
-				warn!("Handler await error: {}", err);
-				let mut ise = Response::default();
-				*ise.status_mut() = StatusCode::INTERNAL_SERVER_ERROR;
-				Ok(ise)
-			}
-			Ok(Err(err)) => {
-				trace!("({}) Request handler failed: {}", path, err);
-				let mut bad_request = Response::new(Body::from(format!("{}", err)));
-				*bad_request.status_mut() = StatusCode::BAD_REQUEST;
-				Ok(bad_request)
-			}
-			Ok(Ok(resp)) => {
-				trace!("({}) Request handler succeeded", path);
-				Ok(resp)
-			}
-		}
-	}
-
-	/// Run the RpcServer
-	pub async fn run(
-		self: Arc<Self>,
-		shutdown_signal: impl Future<Output = ()>,
-	) -> Result<(), Error> {
-		if let Some(tls_config) = self.tls_config.as_ref() {
-			let ca_certs = tls_util::load_certs(&tls_config.ca_cert)?;
-			let node_certs = tls_util::load_certs(&tls_config.node_cert)?;
-			let node_key = tls_util::load_private_key(&tls_config.node_key)?;
-
-			let mut ca_store = rustls::RootCertStore::empty();
-			for crt in ca_certs.iter() {
-				ca_store.add(crt)?;
-			}
-
-			let mut config =
-				rustls::ServerConfig::new(rustls::AllowAnyAuthenticatedClient::new(ca_store));
-			config.set_single_cert([&node_certs[..], &ca_certs[..]].concat(), node_key)?;
-			let tls_acceptor = Arc::new(TlsAcceptor::from(Arc::new(config)));
-
-			let listener = TcpListener::bind(&self.bind_addr).await?;
-			let incoming = TcpListenerStream::new(listener).filter_map(|socket| async {
-				match socket {
-					Ok(stream) => match tls_acceptor.clone().accept(stream).await {
-						Ok(x) => Some(Ok::<_, hyper::Error>(x)),
-						Err(_e) => None,
-					},
-					Err(_) => None,
-				}
-			});
-			let incoming = hyper::server::accept::from_stream(incoming);
-
-			let self_arc = self.clone();
-			let service = make_service_fn(|conn: &TlsStream<TcpStream>| {
-				let client_addr = conn
-					.get_ref()
-					.0
-					.peer_addr()
-					.unwrap_or_else(|_| ([0, 0, 0, 0], 0).into());
-				let self_arc = self_arc.clone();
-				async move {
-					Ok::<_, Error>(service_fn(move |req: Request<Body>| {
-						self_arc.clone().handler(req, client_addr).map_err(|e| {
-							warn!("RPC handler error: {}", e);
-							e
-						})
-					}))
-				}
-			});
-
-			let server = Server::builder(incoming).serve(service);
-
-			let graceful = server.with_graceful_shutdown(shutdown_signal);
-			info!("RPC server listening on http://{}", self.bind_addr);
-
-			graceful.await?;
-		} else {
-			let self_arc = self.clone();
-			let service = make_service_fn(move |conn: &AddrStream| {
-				let client_addr = conn.remote_addr();
-				let self_arc = self_arc.clone();
-				async move {
-					Ok::<_, Error>(service_fn(move |req: Request<Body>| {
-						self_arc.clone().handler(req, client_addr).map_err(|e| {
-							warn!("RPC handler error: {}", e);
-							e
-						})
-					}))
-				}
-			});
-
-			let server = Server::bind(&self.bind_addr).serve(service);
-
-			let graceful = server.with_graceful_shutdown(shutdown_signal);
-			info!("RPC server listening on http://{}", self.bind_addr);
-
-			graceful.await?;
-		}
-
-		Ok(())
-	}
-}

src/rpc/system.rs

@@ -0,0 +1,564 @@
+//! Module containing structs related to membership management
+use std::collections::HashMap;
+use std::io::{Read, Write};
+use std::net::SocketAddr;
+use std::path::Path;
+use std::sync::{Arc, RwLock};
+use std::time::{Duration, Instant};
+
+use arc_swap::ArcSwap;
+use async_trait::async_trait;
+use futures::{join, select};
+use futures_util::future::*;
+use serde::{Deserialize, Serialize};
+use sodiumoxide::crypto::sign::ed25519;
+use tokio::sync::watch;
+use tokio::sync::Mutex;
+
+use netapp::endpoint::{Endpoint, EndpointHandler};
+use netapp::peering::fullmesh::FullMeshPeeringStrategy;
+use netapp::proto::*;
+use netapp::util::parse_and_resolve_peer_addr;
+use netapp::{NetApp, NetworkKey, NodeID, NodeKey};
+
+use garage_util::background::BackgroundRunner;
+use garage_util::config::Config;
+use garage_util::data::Uuid;
+use garage_util::error::*;
+use garage_util::persister::Persister;
+use garage_util::time::*;
+
+use crate::consul::*;
+use crate::ring::*;
+use crate::rpc_helper::*;
+
+const DISCOVERY_INTERVAL: Duration = Duration::from_secs(60);
+const STATUS_EXCHANGE_INTERVAL: Duration = Duration::from_secs(10);
+const PING_TIMEOUT: Duration = Duration::from_secs(2);
+
+/// RPC endpoint used for calls related to membership
+pub const SYSTEM_RPC_PATH: &str = "garage_rpc/membership.rs/SystemRpc";
+
+pub const CONNECT_ERROR_MESSAGE: &str = "Error establishing RPC connection to remote node. This can happen if the remote node is not reachable on the network, but also if the two nodes are not configured with the same rpc_secret";
+
+/// RPC messages related to membership
+#[derive(Debug, Serialize, Deserialize, Clone)]
+pub enum SystemRpc {
+	/// Response to successfull advertisements
+	Ok,
+	/// Request to connect to a specific node (in <pubkey>@<host>:<port> format)
+	Connect(String),
+	/// Ask other node its config. Answered with AdvertiseConfig
+	PullConfig,
+	/// Advertise Garage status. Answered with another AdvertiseStatus.
+	/// Exchanged with every node on a regular basis.
+	AdvertiseStatus(NodeStatus),
+	/// Advertisement of nodes config. Sent spontanously or in response to PullConfig
+	AdvertiseConfig(NetworkConfig),
+	/// Get known nodes states
+	GetKnownNodes,
+	/// Return known nodes
+	ReturnKnownNodes(Vec<KnownNodeInfo>),
+}
+
+impl Rpc for SystemRpc {
+	type Response = Result<SystemRpc, Error>;
+}
+
+/// This node's membership manager
+pub struct System {
+	/// The id of this node
+	pub id: Uuid,
+
+	persist_config: Persister<NetworkConfig>,
+	persist_peer_list: Persister<Vec<(Uuid, SocketAddr)>>,
+
+	local_status: ArcSwap<NodeStatus>,
+	node_status: RwLock<HashMap<Uuid, (u64, NodeStatus)>>,
+
+	pub netapp: Arc<NetApp>,
+	fullmesh: Arc<FullMeshPeeringStrategy>,
+	pub rpc: RpcHelper,
+
+	system_endpoint: Arc<Endpoint<SystemRpc, System>>,
+
+	rpc_listen_addr: SocketAddr,
+	rpc_public_addr: Option<SocketAddr>,
+	bootstrap_peers: Vec<(NodeID, SocketAddr)>,
+	consul_host: Option<String>,
+	consul_service_name: Option<String>,
+	replication_factor: usize,
+
+	/// The ring
+	pub ring: watch::Receiver<Arc<Ring>>,
+	update_ring: Mutex<watch::Sender<Arc<Ring>>>,
+
+	/// The job runner of this node
+	pub background: Arc<BackgroundRunner>,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct NodeStatus {
+	/// Hostname of the node
+	pub hostname: String,
+	/// Replication factor configured on the node
+	pub replication_factor: usize,
+	/// Configuration version
+	pub config_version: u64,
+}
+
+#[derive(Debug, Clone, Serialize, Deserialize)]
+pub struct KnownNodeInfo {
+	pub id: Uuid,
+	pub addr: SocketAddr,
+	pub is_up: bool,
+	pub last_seen_secs_ago: Option<u64>,
+	pub status: NodeStatus,
+}
+
+pub fn read_node_id(metadata_dir: &Path) -> Result<NodeID, Error> {
+	let mut pubkey_file = metadata_dir.to_path_buf();
+	pubkey_file.push("node_key.pub");
+
+	let mut f = std::fs::File::open(pubkey_file.as_path())?;
+	let mut d = vec![];
+	f.read_to_end(&mut d)?;
+	if d.len() != 32 {
+		return Err(Error::Message("Corrupt node_key.pub file".to_string()));
+	}
+
+	let mut key = [0u8; 32];
+	key.copy_from_slice(&d[..]);
+	Ok(NodeID::from_slice(&key[..]).unwrap())
+}
+
+pub fn gen_node_key(metadata_dir: &Path) -> Result<NodeKey, Error> {
+	let mut key_file = metadata_dir.to_path_buf();
+	key_file.push("node_key");
+	if key_file.as_path().exists() {
+		let mut f = std::fs::File::open(key_file.as_path())?;
+		let mut d = vec![];
+		f.read_to_end(&mut d)?;
+		if d.len() != 64 {
+			return Err(Error::Message("Corrupt node_key file".to_string()));
+		}
+
+		let mut key = [0u8; 64];
+		key.copy_from_slice(&d[..]);
+		Ok(NodeKey::from_slice(&key[..]).unwrap())
+	} else {
+		if !metadata_dir.exists() {
+			info!("Metadata directory does not exist, creating it.");
+			std::fs::create_dir(&metadata_dir)?;
+		}
+
+		info!("Generating new node key pair.");
+		let (pubkey, key) = ed25519::gen_keypair();
+
+		{
+			use std::os::unix::fs::PermissionsExt;
+			let mut f = std::fs::File::create(key_file.as_path())?;
+			let mut perm = f.metadata()?.permissions();
+			perm.set_mode(0o600);
+			std::fs::set_permissions(key_file.as_path(), perm)?;
+			f.write_all(&key[..])?;
+		}
+
+		{
+			let mut pubkey_file = metadata_dir.to_path_buf();
+			pubkey_file.push("node_key.pub");
+			let mut f2 = std::fs::File::create(pubkey_file.as_path())?;
+			f2.write_all(&pubkey[..])?;
+		}
+
+		Ok(key)
+	}
+}
+
+impl System {
+	/// Create this node's membership manager
+	pub fn new(
+		network_key: NetworkKey,
+		background: Arc<BackgroundRunner>,
+		replication_factor: usize,
+		config: &Config,
+	) -> Arc<Self> {
+		let node_key =
+			gen_node_key(&config.metadata_dir).expect("Unable to read or generate node ID");
+		info!("Node public key: {}", hex::encode(&node_key.public_key()));
+
+		let persist_config = Persister::new(&config.metadata_dir, "network_config");
+		let persist_peer_list = Persister::new(&config.metadata_dir, "peer_list");
+
+		let net_config = match persist_config.load() {
+			Ok(x) => x,
+			Err(e) => {
+				info!(
+					"No valid previous network configuration stored ({}), starting fresh.",
+					e
+				);
+				NetworkConfig::new()
+			}
+		};
+
+		let local_status = NodeStatus {
+			hostname: gethostname::gethostname()
+				.into_string()
+				.unwrap_or_else(|_| "<invalid utf-8>".to_string()),
+			replication_factor,
+			config_version: net_config.version,
+		};
+
+		let ring = Ring::new(net_config, replication_factor);
+		let (update_ring, ring) = watch::channel(Arc::new(ring));
+
+		if let Some(addr) = config.rpc_public_addr {
+			println!("{}@{}", hex::encode(&node_key.public_key()), addr);
+		} else {
+			println!("{}", hex::encode(&node_key.public_key()));
+		}
+
+		let netapp = NetApp::new(network_key, node_key);
+		let fullmesh = FullMeshPeeringStrategy::new(
+			netapp.clone(),
+			config.bootstrap_peers.clone(),
+			config.rpc_public_addr,
+		);
+
+		let system_endpoint = netapp.endpoint(SYSTEM_RPC_PATH.into());
+
+		let sys = Arc::new(System {
+			id: netapp.id.into(),
+			persist_config,
+			persist_peer_list,
+			local_status: ArcSwap::new(Arc::new(local_status)),
+			node_status: RwLock::new(HashMap::new()),
+			netapp: netapp.clone(),
+			fullmesh: fullmesh.clone(),
+			rpc: RpcHelper::new(netapp.id.into(), fullmesh, background.clone(), ring.clone()),
+			system_endpoint,
+			replication_factor,
+			rpc_listen_addr: config.rpc_bind_addr,
+			rpc_public_addr: config.rpc_public_addr,
+			bootstrap_peers: config.bootstrap_peers.clone(),
+			consul_host: config.consul_host.clone(),
+			consul_service_name: config.consul_service_name.clone(),
+			ring,
+			update_ring: Mutex::new(update_ring),
+			background,
+		});
+		sys.system_endpoint.set_handler(sys.clone());
+		sys
+	}
+
+	/// Perform bootstraping, starting the ping loop
+	pub async fn run(self: Arc<Self>, must_exit: watch::Receiver<bool>) {
+		join!(
+			self.netapp
+				.clone()
+				.listen(self.rpc_listen_addr, None, must_exit.clone()),
+			self.fullmesh.clone().run(must_exit.clone()),
+			self.discovery_loop(must_exit.clone()),
+			self.status_exchange_loop(must_exit.clone()),
+		);
+	}
+
+	// ---- INTERNALS ----
+
+	async fn advertise_to_consul(self: Arc<Self>) -> Result<(), Error> {
+		let (consul_host, consul_service_name) =
+			match (&self.consul_host, &self.consul_service_name) {
+				(Some(ch), Some(csn)) => (ch, csn),
+				_ => return Ok(()),
+			};
+
+		let rpc_public_addr = match self.rpc_public_addr {
+			Some(addr) => addr,
+			None => {
+				warn!("Not advertising to Consul because rpc_public_addr is not defined in config file.");
+				return Ok(());
+			}
+		};
+
+		publish_consul_service(
+			consul_host,
+			consul_service_name,
+			self.netapp.id,
+			&self.local_status.load_full().hostname,
+			rpc_public_addr,
+		)
+		.await
+		.err_context("Error while publishing Consul service")
+	}
+
+	/// Save network configuration to disc
+	async fn save_network_config(self: Arc<Self>) -> Result<(), Error> {
+		let ring: Arc<Ring> = self.ring.borrow().clone();
+		self.persist_config
+			.save_async(&ring.config)
+			.await
+			.expect("Cannot save current cluster configuration");
+		Ok(())
+	}
+
+	fn update_local_status(&self) {
+		let mut new_si: NodeStatus = self.local_status.load().as_ref().clone();
+
+		let ring = self.ring.borrow();
+		new_si.config_version = ring.config.version;
+		self.local_status.swap(Arc::new(new_si));
+	}
+
+	async fn handle_connect(&self, node: &str) -> Result<SystemRpc, Error> {
+		let (pubkey, addrs) = parse_and_resolve_peer_addr(node).ok_or_else(|| {
+			Error::Message(format!(
+				"Unable to parse or resolve node specification: {}",
+				node
+			))
+		})?;
+		let mut errors = vec![];
+		for ip in addrs.iter() {
+			match self
+				.netapp
+				.clone()
+				.try_connect(*ip, pubkey)
+				.await
+				.err_context(CONNECT_ERROR_MESSAGE)
+			{
+				Ok(()) => return Ok(SystemRpc::Ok),
+				Err(e) => {
+					errors.push((*ip, e));
+				}
+			}
+		}
+		return Err(Error::Message(format!(
+			"Could not connect to specified peers. Errors: {:?}",
+			errors
+		)));
+	}
+
+	fn handle_pull_config(&self) -> SystemRpc {
+		let ring = self.ring.borrow().clone();
+		SystemRpc::AdvertiseConfig(ring.config.clone())
+	}
+
+	fn handle_get_known_nodes(&self) -> SystemRpc {
+		let node_status = self.node_status.read().unwrap();
+		let known_nodes = self
+			.fullmesh
+			.get_peer_list()
+			.iter()
+			.map(|n| KnownNodeInfo {
+				id: n.id.into(),
+				addr: n.addr,
+				is_up: n.is_up(),
+				last_seen_secs_ago: n.last_seen.map(|t| (Instant::now() - t).as_secs()),
+				status: node_status
+					.get(&n.id.into())
+					.cloned()
+					.map(|(_, st)| st)
+					.unwrap_or(NodeStatus {
+						hostname: "?".to_string(),
+						replication_factor: 0,
+						config_version: 0,
+					}),
+			})
+			.collect::<Vec<_>>();
+		SystemRpc::ReturnKnownNodes(known_nodes)
+	}
+
+	async fn handle_advertise_status(
+		self: &Arc<Self>,
+		from: Uuid,
+		info: &NodeStatus,
+	) -> Result<SystemRpc, Error> {
+		let local_info = self.local_status.load();
+
+		if local_info.replication_factor < info.replication_factor {
+			error!("Some node have a higher replication factor ({}) than this one ({}). This is not supported and might lead to bugs",
+				info.replication_factor,
+				local_info.replication_factor);
+			std::process::exit(1);
+		}
+
+		if info.config_version > local_info.config_version {
+			let self2 = self.clone();
+			self.background.spawn_cancellable(async move {
+				self2.pull_config(from).await;
+				Ok(())
+			});
+		}
+
+		self.node_status
+			.write()
+			.unwrap()
+			.insert(from, (now_msec(), info.clone()));
+
+		Ok(SystemRpc::Ok)
+	}
+
+	async fn handle_advertise_config(
+		self: Arc<Self>,
+		adv: &NetworkConfig,
+	) -> Result<SystemRpc, Error> {
+		let update_ring = self.update_ring.lock().await;
+		let ring: Arc<Ring> = self.ring.borrow().clone();
+
+		if adv.version > ring.config.version {
+			let ring = Ring::new(adv.clone(), self.replication_factor);
+			update_ring.send(Arc::new(ring))?;
+			drop(update_ring);
+
+			let self2 = self.clone();
+			let adv = adv.clone();
+			self.background.spawn_cancellable(async move {
+				self2
+					.rpc
+					.broadcast(
+						&self2.system_endpoint,
+						SystemRpc::AdvertiseConfig(adv),
+						RequestStrategy::with_priority(PRIO_HIGH),
+					)
+					.await;
+				Ok(())
+			});
+			self.background.spawn(self.clone().save_network_config());
+		}
+
+		Ok(SystemRpc::Ok)
+	}
+
+	async fn status_exchange_loop(&self, mut stop_signal: watch::Receiver<bool>) {
+		while !*stop_signal.borrow() {
+			let restart_at = tokio::time::sleep(STATUS_EXCHANGE_INTERVAL);
+
+			self.update_local_status();
+			let local_status: NodeStatus = self.local_status.load().as_ref().clone();
+			self.rpc
+				.broadcast(
+					&self.system_endpoint,
+					SystemRpc::AdvertiseStatus(local_status),
+					RequestStrategy::with_priority(PRIO_HIGH).with_timeout(PING_TIMEOUT),
+				)
+				.await;
+
+			select! {
+				_ = restart_at.fuse() => {},
+				_ = stop_signal.changed().fuse() => {},
+			}
+		}
+	}
+
+	async fn discovery_loop(self: &Arc<Self>, mut stop_signal: watch::Receiver<bool>) {
+		let consul_config = match (&self.consul_host, &self.consul_service_name) {
+			(Some(ch), Some(csn)) => Some((ch.clone(), csn.clone())),
+			_ => None,
+		};
+
+		while !*stop_signal.borrow() {
+			let not_configured = self.ring.borrow().config.members.is_empty();
+			let no_peers = self.fullmesh.get_peer_list().len() < self.replication_factor;
+			let bad_peers = self
+				.fullmesh
+				.get_peer_list()
+				.iter()
+				.filter(|p| p.is_up())
+				.count() != self.ring.borrow().config.members.len();
+
+			if not_configured || no_peers || bad_peers {
+				info!("Doing a bootstrap/discovery step (not_configured: {}, no_peers: {}, bad_peers: {})", not_configured, no_peers, bad_peers);
+
+				let mut ping_list = self.bootstrap_peers.clone();
+
+				// Add peer list from list stored on disk
+				if let Ok(peers) = self.persist_peer_list.load_async().await {
+					ping_list.extend(peers.iter().map(|(id, addr)| ((*id).into(), *addr)))
+				}
+
+				// Fetch peer list from Consul
+				if let Some((consul_host, consul_service_name)) = &consul_config {
+					match get_consul_nodes(consul_host, consul_service_name).await {
+						Ok(node_list) => {
+							ping_list.extend(node_list);
+						}
+						Err(e) => {
+							warn!("Could not retrieve node list from Consul: {}", e);
+						}
+					}
+				}
+
+				for (node_id, node_addr) in ping_list {
+					tokio::spawn(
+						self.netapp
+							.clone()
+							.try_connect(node_addr, node_id)
+							.map(|r| r.err_context(CONNECT_ERROR_MESSAGE)),
+					);
+				}
+			}
+
+			if let Err(e) = self.save_peer_list().await {
+				warn!("Could not save peer list to file: {}", e);
+			}
+
+			self.background.spawn(self.clone().advertise_to_consul());
+
+			let restart_at = tokio::time::sleep(DISCOVERY_INTERVAL);
+			select! {
+				_ = restart_at.fuse() => {},
+				_ = stop_signal.changed().fuse() => {},
+			}
+		}
+	}
+
+	async fn save_peer_list(&self) -> Result<(), Error> {
+		// Prepare new peer list to save to file
+		// It is a vec of tuples (node ID as Uuid, node SocketAddr)
+		let mut peer_list = self
+			.fullmesh
+			.get_peer_list()
+			.iter()
+			.map(|n| (n.id.into(), n.addr))
+			.collect::<Vec<_>>();
+
+		// Before doing it, we read the current peer list file (if it exists)
+		// and append it to the list we are about to save,
+		// so that no peer ID gets lost in the process.
+		if let Ok(mut prev_peer_list) = self.persist_peer_list.load_async().await {
+			prev_peer_list.retain(|(id, _ip)| peer_list.iter().all(|(id2, _ip2)| id2 != id));
+			peer_list.extend(prev_peer_list);
+		}
+
+		// Save new peer list to file
+		self.persist_peer_list.save_async(&peer_list).await
+	}
+
+	async fn pull_config(self: Arc<Self>, peer: Uuid) {
+		let resp = self
+			.rpc
+			.call(
+				&self.system_endpoint,
+				peer,
+				SystemRpc::PullConfig,
+				RequestStrategy::with_priority(PRIO_HIGH).with_timeout(PING_TIMEOUT),
+			)
+			.await;
+		if let Ok(SystemRpc::AdvertiseConfig(config)) = resp {
+			let _: Result<_, _> = self.handle_advertise_config(&config).await;
+		}
+	}
+}
+
+#[async_trait]
+impl EndpointHandler<SystemRpc> for System {
+	async fn handle(self: &Arc<Self>, msg: &SystemRpc, from: NodeID) -> Result<SystemRpc, Error> {
+		match msg {
+			SystemRpc::Connect(node) => self.handle_connect(node).await,
+			SystemRpc::PullConfig => Ok(self.handle_pull_config()),
+			SystemRpc::AdvertiseStatus(adv) => self.handle_advertise_status(from.into(), adv).await,
+			SystemRpc::AdvertiseConfig(adv) => self.clone().handle_advertise_config(adv).await,
+			SystemRpc::GetKnownNodes => Ok(self.handle_get_known_nodes()),
+			_ => Err(Error::BadRpc("Unexpected RPC message".to_string())),
+		}
+	}
+}

src/rpc/tls_util.rs

@@ -1,139 +0,0 @@
-use core::future::Future;
-use core::task::{Context, Poll};
-use std::pin::Pin;
-use std::sync::Arc;
-use std::{fs, io};
-
-use futures_util::future::*;
-use hyper::client::connect::Connection;
-use hyper::client::HttpConnector;
-use hyper::service::Service;
-use hyper::Uri;
-use hyper_rustls::MaybeHttpsStream;
-use rustls::internal::pemfile;
-use tokio::io::{AsyncRead, AsyncWrite};
-use tokio_rustls::TlsConnector;
-use webpki::DNSNameRef;
-
-use garage_util::error::Error;
-
-pub fn load_certs(filename: &str) -> Result<Vec<rustls::Certificate>, Error> {
-	let certfile = fs::File::open(&filename)?;
-	let mut reader = io::BufReader::new(certfile);
-
-	let certs = pemfile::certs(&mut reader).map_err(|_| {
-		Error::Message(format!(
-			"Could not deecode certificates from file: {}",
-			filename
-		))
-	})?;
-
-	if certs.is_empty() {
-		return Err(Error::Message(format!(
-			"Invalid certificate file: {}",
-			filename
-		)));
-	}
-	Ok(certs)
-}
-
-pub fn load_private_key(filename: &str) -> Result<rustls::PrivateKey, Error> {
-	let keyfile = fs::File::open(&filename)?;
-	let mut reader = io::BufReader::new(keyfile);
-
-	let keys = pemfile::rsa_private_keys(&mut reader).map_err(|_| {
-		Error::Message(format!(
-			"Could not decode private key from file: {}",
-			filename
-		))
-	})?;
-
-	if keys.len() != 1 {
-		return Err(Error::Message(format!(
-			"Invalid private key file: {} ({} private keys)",
-			filename,
-			keys.len()
-		)));
-	}
-	Ok(keys[0].clone())
-}
-
-// ---- AWFUL COPYPASTA FROM HYPER-RUSTLS connector.rs
-// ---- ALWAYS USE `garage` AS HOSTNAME FOR TLS VERIFICATION
-
-#[derive(Clone)]
-pub struct HttpsConnectorFixedDnsname<T> {
-	http: T,
-	tls_config: Arc<rustls::ClientConfig>,
-	fixed_dnsname: &'static str,
-}
-
-type BoxError = Box<dyn std::error::Error + Send + Sync>;
-
-impl HttpsConnectorFixedDnsname<HttpConnector> {
-	pub fn new(mut tls_config: rustls::ClientConfig, fixed_dnsname: &'static str) -> Self {
-		let mut http = HttpConnector::new();
-		http.enforce_http(false);
-		tls_config.alpn_protocols = vec![b"h2".to_vec(), b"http/1.1".to_vec()];
-		Self {
-			http,
-			tls_config: Arc::new(tls_config),
-			fixed_dnsname,
-		}
-	}
-}
-
-impl<T> Service<Uri> for HttpsConnectorFixedDnsname<T>
-where
-	T: Service<Uri>,
-	T::Response: Connection + AsyncRead + AsyncWrite + Send + Unpin + 'static,
-	T::Future: Send + 'static,
-	T::Error: Into<BoxError>,
-{
-	type Response = MaybeHttpsStream<T::Response>;
-	type Error = BoxError;
-
-	#[allow(clippy::type_complexity)]
-	type Future =
-		Pin<Box<dyn Future<Output = Result<MaybeHttpsStream<T::Response>, BoxError>> + Send>>;
-
-	fn poll_ready(&mut self, cx: &mut Context<'_>) -> Poll<Result<(), Self::Error>> {
-		match self.http.poll_ready(cx) {
-			Poll::Ready(Ok(())) => Poll::Ready(Ok(())),
-			Poll::Ready(Err(e)) => Poll::Ready(Err(e.into())),
-			Poll::Pending => Poll::Pending,
-		}
-	}
-
-	fn call(&mut self, dst: Uri) -> Self::Future {
-		let is_https = dst.scheme_str() == Some("https");
-
-		if !is_https {
-			let connecting_future = self.http.call(dst);
-
-			let f = async move {
-				let tcp = connecting_future.await.map_err(Into::into)?;
-
-				Ok(MaybeHttpsStream::Http(tcp))
-			};
-			f.boxed()
-		} else {
-			let cfg = self.tls_config.clone();
-			let connecting_future = self.http.call(dst);
-
-			let dnsname =
-				DNSNameRef::try_from_ascii_str(self.fixed_dnsname).expect("Invalid fixed dnsname");
-
-			let f = async move {
-				let tcp = connecting_future.await.map_err(Into::into)?;
-				let connector = TlsConnector::from(cfg);
-				let tls = connector
-					.connect(dnsname, tcp)
-					.await
-					.map_err(|e| io::Error::new(io::ErrorKind::Other, e))?;
-				Ok(MaybeHttpsStream::Https(tls))
-			};
-			f.boxed()
-		}
-	}
-}

src/table/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "garage_table"
-version = "0.3.0"
+version = "0.4.0"
 authors = ["Alex Auvolat <alex@adnab.me>"]
 edition = "2018"
 license = "AGPL-3.0"
@@ -13,9 +13,10 @@ path = "lib.rs"
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [dependencies]
-garage_rpc = { version = "0.3.0", path = "../rpc" }
-garage_util = { version = "0.3.0", path = "../util" }
+garage_rpc = { version = "0.4.0", path = "../rpc" }
+garage_util = { version = "0.4.0", path = "../util" }
 
+async-trait = "0.1.7"
 bytes = "1.0"
 hexdump = "0.1"
 log = "0.4"
@@ -30,4 +31,3 @@ serde_bytes = "0.11"
 futures = "0.3"
 futures-util = "0.3"
 tokio = { version = "1.0", default-features = false, features = ["rt", "rt-multi-thread", "io-util", "net", "time", "macros", "sync", "signal", "fs"] }
-

src/table/crdt/lww_map.rs

@@ -103,7 +103,7 @@ where
 	}
 	/// Get a reference to the value assigned to a key
 	pub fn get(&self, k: &K) -> Option<&V> {
-		match self.vals.binary_search_by(|(k2, _, _)| k2.cmp(&k)) {
+		match self.vals.binary_search_by(|(k2, _, _)| k2.cmp(k)) {
 			Ok(i) => Some(&self.vals[i].2),
 			Err(_) => None,
 		}
@@ -132,14 +132,14 @@ where
 {
 	fn merge(&mut self, other: &Self) {
 		for (k, ts2, v2) in other.vals.iter() {
-			match self.vals.binary_search_by(|(k2, _, _)| k2.cmp(&k)) {
+			match self.vals.binary_search_by(|(k2, _, _)| k2.cmp(k)) {
 				Ok(i) => {
 					let (_, ts1, _v1) = &self.vals[i];
 					if ts2 > ts1 {
 						self.vals[i].1 = *ts2;
 						self.vals[i].2 = v2.clone();
 					} else if ts1 == ts2 {
-						self.vals[i].2.merge(&v2);
+						self.vals[i].2.merge(v2);
 					}
 				}
 				Err(i) => {

src/table/crdt/map.rs

@@ -49,7 +49,7 @@ where
 
 	/// Get a reference to the value assigned to a key
 	pub fn get(&self, k: &K) -> Option<&V> {
-		match self.vals.binary_search_by(|(k2, _)| k2.cmp(&k)) {
+		match self.vals.binary_search_by(|(k2, _)| k2.cmp(k)) {
 			Ok(i) => Some(&self.vals[i].1),
 			Err(_) => None,
 		}
@@ -76,9 +76,9 @@ where
 {
 	fn merge(&mut self, other: &Self) {
 		for (k, v2) in other.vals.iter() {
-			match self.vals.binary_search_by(|(k2, _)| k2.cmp(&k)) {
+			match self.vals.binary_search_by(|(k2, _)| k2.cmp(k)) {
 				Ok(i) => {
-					self.vals[i].1.merge(&v2);
+					self.vals[i].1.merge(v2);
 				}
 				Err(i) => {
 					self.vals.insert(i, (k.clone(), v2.clone()));

src/table/data.rs

@@ -9,7 +9,7 @@ use tokio::sync::Notify;
 use garage_util::data::*;
 use garage_util::error::*;
 
-use garage_rpc::membership::System;
+use garage_rpc::system::System;
 
 use crate::crdt::Crdt;
 use crate::replication::*;

src/table/gc.rs

@@ -2,6 +2,7 @@ use std::collections::HashMap;
 use std::sync::Arc;
 use std::time::Duration;
 
+use async_trait::async_trait;
 use serde::{Deserialize, Serialize};
 use serde_bytes::ByteBuf;
 
@@ -13,9 +14,8 @@ use tokio::sync::watch;
 use garage_util::data::*;
 use garage_util::error::Error;
 
-use garage_rpc::membership::System;
-use garage_rpc::rpc_client::*;
-use garage_rpc::rpc_server::*;
+use garage_rpc::system::System;
+use garage_rpc::*;
 
 use crate::data::*;
 use crate::replication::*;
@@ -24,11 +24,11 @@ use crate::schema::*;
 const TABLE_GC_BATCH_SIZE: usize = 1024;
 const TABLE_GC_RPC_TIMEOUT: Duration = Duration::from_secs(30);
 
-pub struct TableGc<F: TableSchema, R: TableReplication> {
+pub struct TableGc<F: TableSchema + 'static, R: TableReplication + 'static> {
 	system: Arc<System>,
 	data: Arc<TableData<F, R>>,
 
-	rpc_client: Arc<RpcClient<GcRpc>>,
+	endpoint: Arc<Endpoint<GcRpc, Self>>,
 }
 
 #[derive(Serialize, Deserialize)]
@@ -38,28 +38,27 @@ enum GcRpc {
 	Ok,
 }
 
-impl RpcMessage for GcRpc {}
+impl Rpc for GcRpc {
+	type Response = Result<GcRpc, Error>;
+}
 
 impl<F, R> TableGc<F, R>
 where
 	F: TableSchema + 'static,
 	R: TableReplication + 'static,
 {
-	pub(crate) fn launch(
-		system: Arc<System>,
-		data: Arc<TableData<F, R>>,
-		rpc_server: &mut RpcServer,
-	) -> Arc<Self> {
-		let rpc_path = format!("table_{}/gc", data.name);
-		let rpc_client = system.rpc_client::<GcRpc>(&rpc_path);
+	pub(crate) fn launch(system: Arc<System>, data: Arc<TableData<F, R>>) -> Arc<Self> {
+		let endpoint = system
+			.netapp
+			.endpoint(format!("garage_table/gc.rs/Rpc:{}", data.name));
 
 		let gc = Arc::new(Self {
 			system: system.clone(),
 			data: data.clone(),
-			rpc_client,
+			endpoint,
 		});
 
-		gc.register_handler(rpc_server, rpc_path);
+		gc.endpoint.set_handler(gc.clone());
 
 		let gc1 = gc.clone();
 		system.background.spawn_worker(
@@ -180,11 +179,15 @@ where
 			deletes.push((k, vhash));
 		}
 
-		self.rpc_client
+		self.system
+			.rpc
 			.try_call_many(
+				&self.endpoint,
 				&nodes[..],
 				GcRpc::Update(updates),
-				RequestStrategy::with_quorum(nodes.len()).with_timeout(TABLE_GC_RPC_TIMEOUT),
+				RequestStrategy::with_priority(PRIO_BACKGROUND)
+					.with_quorum(nodes.len())
+					.with_timeout(TABLE_GC_RPC_TIMEOUT),
 			)
 			.await?;
 
@@ -193,11 +196,15 @@ where
 			self.data.name, n_items
 		);
 
-		self.rpc_client
+		self.system
+			.rpc
 			.try_call_many(
+				&self.endpoint,
 				&nodes[..],
 				GcRpc::DeleteIfEqualHash(deletes.clone()),
-				RequestStrategy::with_quorum(nodes.len()).with_timeout(TABLE_GC_RPC_TIMEOUT),
+				RequestStrategy::with_priority(PRIO_BACKGROUND)
+					.with_quorum(nodes.len())
+					.with_timeout(TABLE_GC_RPC_TIMEOUT),
 			)
 			.await?;
 
@@ -216,25 +223,15 @@ where
 			.compare_and_swap::<_, _, Vec<u8>>(key, Some(vhash), None)?;
 		Ok(())
 	}
+}
 
-	// ---- RPC HANDLER ----
-
-	fn register_handler(self: &Arc<Self>, rpc_server: &mut RpcServer, path: String) {
-		let self2 = self.clone();
-		rpc_server.add_handler::<GcRpc, _, _>(path, move |msg, _addr| {
-			let self2 = self2.clone();
-			async move { self2.handle_rpc(&msg).await }
-		});
-
-		let self2 = self.clone();
-		self.rpc_client
-			.set_local_handler(self.system.id, move |msg| {
-				let self2 = self2.clone();
-				async move { self2.handle_rpc(&msg).await }
-			});
-	}
-
-	async fn handle_rpc(self: &Arc<Self>, message: &GcRpc) -> Result<GcRpc, Error> {
+#[async_trait]
+impl<F, R> EndpointHandler<GcRpc> for TableGc<F, R>
+where
+	F: TableSchema + 'static,
+	R: TableReplication + 'static,
+{
+	async fn handle(self: &Arc<Self>, message: &GcRpc, _from: NodeID) -> Result<GcRpc, Error> {
 		match message {
 			GcRpc::Update(items) => {
 				self.data.update_many(items)?;

src/table/merkle.rs

@@ -167,7 +167,7 @@ where
 		// Calculate an update to apply to this node
 		// This update is an Option<_>, so that it is None if the update is a no-op
 		// and we can thus skip recalculating and re-storing everything
-		let mutate = match self.read_node_txn(tx, &key)? {
+		let mutate = match self.read_node_txn(tx, key)? {
 			MerkleNode::Empty => new_vhash.map(|vhv| MerkleNode::Leaf(k.to_vec(), vhv)),
 			MerkleNode::Intermediate(mut children) => {
 				let key2 = key.next_key(khash);
@@ -270,7 +270,7 @@ where
 		};
 
 		if let Some(new_node) = mutate {
-			let hash = self.put_node_txn(tx, &key, &new_node)?;
+			let hash = self.put_node_txn(tx, key, &new_node)?;
 			Ok(Some(hash))
 		} else {
 			Ok(None)