Compare commits

...

24 Commits

Author SHA1 Message Date
Alex Auvolat c705a8f66e Enable k2v feature flag in CI 2022-07-18 12:19:18 +02:00
Alex Auvolat ac03fa7937 Uniformize tracing::* imports (hopefully fixes 32-bit build) 2022-07-15 18:31:19 +02:00
Alex 4f38cadf6e Background task manager (#332)
- [x] New background worker trait
- [x] Adapt all current workers to use new API
- [x] Command to list currently running workers, and whether they are active, idle, or dead
- [x] Error reporting
- Optimizations
  - [x] Merkle updater: several items per iteration
  - [ ] Use `tokio::task::spawn_blocking` where appropriate so that CPU-intensive tasks don't block other things going on
- scrub:
  - [x] have only one worker with a channel to start/pause/cancel
  - [x] automatic scrub
  - [x] ability to view and change tranquility from CLI
  - [x] persistence of a few info
- [ ] Testing

Co-authored-by: Alex Auvolat <alex@adnab.me>
Reviewed-on: https://git.deuxfleurs.fr/Deuxfleurs/garage/pulls/332
Co-authored-by: Alex <alex@adnab.me>
Co-committed-by: Alex <alex@adnab.me>
2022-07-08 13:30:26 +02:00
Alex Auvolat aab34bfe54 add delays in k2v test_items_and_indices 2022-07-08 10:41:57 +02:00
Alex fe3fa83de7 Publish k2v-client crate to crates.io (#337)
Co-authored-by: Alex Auvolat <alex@adnab.me>
Reviewed-on: https://git.deuxfleurs.fr/Deuxfleurs/garage/pulls/337
Co-authored-by: Alex <alex@adnab.me>
Co-committed-by: Alex <alex@adnab.me>
2022-07-04 18:27:25 +02:00
Alex Auvolat b6d59ec19a Fix poll item when item didn't change 2022-07-04 14:00:02 +02:00
Alex 0850bac874 Add poll command to k2v-cli (#335)
Co-authored-by: Alex Auvolat <alex@adnab.me>
Reviewed-on: https://git.deuxfleurs.fr/Deuxfleurs/garage/pulls/335
Co-authored-by: Alex <alex@adnab.me>
Co-committed-by: Alex <alex@adnab.me>
2022-07-04 12:45:32 +02:00
Alex b74b533b7b Fix typo 2022-06-29 11:50:51 +02:00
Alex 996f2a6d58 Slides for talk at IMT Atlantique / STACK on 2022-06-23 (#333)
Co-authored-by: Alex Auvolat <alex@adnab.me>
Reviewed-on: https://git.deuxfleurs.fr/Deuxfleurs/garage/pulls/333
Co-authored-by: Alex <alex@adnab.me>
Co-committed-by: Alex <alex@adnab.me>
2022-06-23 14:28:40 +02:00
Alex 77e3fd6db2 improve internal item counter mechanisms and implement bucket quotas (#326)
- [x] Refactoring of internal counting API
- [x] Repair procedure for counters (it's an offline procedure!!!)
- [x] New counter for objects in buckets
- [x] Add quotas to buckets struct
- [x] Add CLI to manage bucket quotas
- [x] Add admin API to manage bucket quotas
- [x] Apply quotas by adding checks on put operations
- [x] Proof-read

Co-authored-by: Alex Auvolat <alex@adnab.me>
Reviewed-on: https://git.deuxfleurs.fr/Deuxfleurs/garage/pulls/326
Co-authored-by: Alex <alex@adnab.me>
Co-committed-by: Alex <alex@adnab.me>
2022-06-15 20:20:28 +02:00
Quentin Dufour d544a0e0e0 Send CORS headers for all requests 2022-06-13 10:19:52 +02:00
Alex Auvolat 138e13071b Fix garage_db build on 32-bit systems 2022-06-09 14:55:20 +02:00
Alex b44d3fc796 Abstract database behind generic interface and implement alternative drivers (#322)
- [x] Design interface
- [x] Implement Sled backend
  - [x] Re-implement the SledCountedTree hack ~~on Sled backend~~ on all backends (i.e. over the abstraction)
- [x] Convert Garage code to use generic interface
- [x] Proof-read converted Garage code
- [ ] Test everything well
- [x] Implement sqlite backend
- [x] Implement LMDB backend
- [ ] (Implement Persy backend?)
- [ ] (Implement other backends? (like RocksDB, ...))
- [x] Implement backend choice in config file and garage server module
- [x] Add CLI for converting between DB formats
- Exploit the new interface to put more things in transactions
  - [x] `.updated()` trigger on Garage tables

Fix #284

**Bugs**

- [x] When exporting sqlite, trees iterate empty??
- [x] LMDB doesn't work

**Known issues for various back-ends**

- Sled:
  - Eats all my RAM and also all my disk space
  - `.len()` has to traverse the whole table
  - Is actually quite slow on some operations
  - And is actually pretty bad code...
- Sqlite:
  - Requires a lock to be taken on all operations. The lock is also taken when iterating on a table with `.iter()`, and the lock isn't released until the iterator is dropped. This means that we must be VERY carefull to not do anything else inside a `.iter()` loop or else we will have a deadlock! Most such cases have been eliminated from the Garage codebase, but there might still be some that remain. If your Garage-over-Sqlite seems to hang/freeze, this is the reason.
  - (adapter uses a bunch of unsafe code)
- Heed (LMDB):
  - Not suited for 32-bit machines as it has to map the whole DB in memory.
  - (adpater uses a tiny bit of unsafe code)

**My recommendation:** avoid 32-bit machines and use LMDB as much as possible.

**Converting databases** is actually quite easy. For example from Sled to LMDB:

```bash
cd src/db
cargo run --features cli --bin convert -- -i path/to/garage/meta/db -a sled -o path/to/garage/meta/db.lmdb -b lmdb
```

Then, just add this to your `config.toml`:

```toml
db_engine = "lmdb"
```

Co-authored-by: Alex Auvolat <alex@adnab.me>
Reviewed-on: https://git.deuxfleurs.fr/Deuxfleurs/garage/pulls/322
Co-authored-by: Alex <alex@adnab.me>
Co-committed-by: Alex <alex@adnab.me>
2022-06-08 10:01:44 +02:00
Simon C 7eed3ceda9 docs: Add Trafik reverse proxy documentation 2022-06-07 16:16:52 +02:00
Simon C 4b8f48f3c5 docs: Fix title level 2022-06-07 13:32:52 +02:00
Simon C 7d3b5585f1 docs: Add link to facilitate navigation in the documentation 2022-06-07 09:38:59 +02:00
Quentin Dufour a1abed0378 Remove useless MC_REGION env variable 2022-06-02 12:50:11 +02:00
Alex Auvolat b54a938724 Fix garage_version() now that GIT_VERSION is read in crate garage_rpc 2022-06-02 12:00:10 +02:00
Alex Auvolat ff06d3f082 Fix Content-Type headers for {admin,k2v} errors and admin responses
Fix #315
2022-05-25 17:09:33 +02:00
Alex 93eab8eaa3 Fixes to S3 compatibility page (#314)
Mention PostObject is implemented, fix english mistakes

Co-authored-by: Alex Auvolat <alex@adnab.me>
Reviewed-on: https://git.deuxfleurs.fr/Deuxfleurs/garage/pulls/314
Co-authored-by: Alex <alex@adnab.me>
Co-committed-by: Alex <alex@adnab.me>
2022-05-25 16:54:44 +02:00
Quentin Dufour 43ddc933f9 Update Ceph S3 endpoints compatibility 2022-05-25 15:20:08 +02:00
Alex Auvolat 9f303f6308 Shorter page title 2022-05-24 15:47:42 +02:00
Alex Auvolat 3be43f3372 Add lost content for Restic with Garage
Suggested-by: Quentin <quentin@deuxfleurs.fr>
2022-05-24 15:32:42 +02:00
Alex Auvolat 2da448b43f Add documentation for new Admin API and a few infos on K2V 2022-05-24 15:28:37 +02:00
134 changed files with 14982 additions and 1732 deletions
Generated
+422 -35
View File
@@ -2,6 +2,17 @@
# It is not intended for manual editing.
version = 3
[[package]]
name = "ahash"
version = "0.7.6"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "fcb51a0695d8f838b1ee009b3fbf66bda078cd64590202a864a8f3e8c4315c47"
dependencies = [
"getrandom",
"once_cell",
"version_check",
]
[[package]]
name = "aho-corasick"
version = "0.7.18"
@@ -82,6 +93,15 @@ dependencies = [
"winapi",
]
[[package]]
name = "autocfg"
version = "0.1.8"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "0dde43e75fd43e8a1bf86103336bc699aa8d17ad1be60c76c0bdfd4828e19b78"
dependencies = [
"autocfg 1.1.0",
]
[[package]]
name = "autocfg"
version = "1.1.0"
@@ -301,6 +321,15 @@ version = "0.13.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "904dfeac50f3cdaba28fc6f57fdcddb75f49ed61346676a78c4ffe55877802fd"
[[package]]
name = "bincode"
version = "1.3.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "b1f45e9417d87227c7a56d22e471c6206462cba514c7590c09aff4cf6d1ddcad"
dependencies = [
"serde",
]
[[package]]
name = "bitflags"
version = "1.3.2"
@@ -333,6 +362,12 @@ version = "3.9.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "a4a45a46ab1f2412e53d3a0ade76ffad2025804294569aae387231a0cd6e0899"
[[package]]
name = "bytemuck"
version = "1.9.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "cdead85bdec19c194affaeeb670c0e41fe23de31459efd1c174d049269cf02cc"
[[package]]
name = "byteorder"
version = "1.4.3"
@@ -361,6 +396,12 @@ dependencies = [
"either",
]
[[package]]
name = "bytesize"
version = "1.1.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "6c58ec36aac5066d5ca17df51b3e70279f5670a72102f5752cb7e7c856adfc70"
[[package]]
name = "cc"
version = "1.0.73"
@@ -370,6 +411,12 @@ dependencies = [
"jobserver",
]
[[package]]
name = "cfg-if"
version = "0.1.10"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "4785bdd1c96b2a846b2bd7cc02e86b6b3dbf14e7e53446c4f54c92a361040822"
[[package]]
name = "cfg-if"
version = "1.0.0"
@@ -486,7 +533,7 @@ version = "1.3.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "b540bd8bc810d3885c6ea91e2018302f68baba2129ab3e88f32389ee9370880d"
dependencies = [
"cfg-if",
"cfg-if 1.0.0",
]
[[package]]
@@ -495,8 +542,8 @@ version = "0.5.4"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "5aaa7bd5fb665c6864b5f963dd9097905c54125909c7aa94c9e18507cdbe6c53"
dependencies = [
"cfg-if",
"crossbeam-utils",
"cfg-if 1.0.0",
"crossbeam-utils 0.8.8",
]
[[package]]
@@ -505,21 +552,40 @@ version = "0.9.8"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "1145cf131a2c6ba0615079ab6a638f7e1973ac9c2634fcbeaaad6114246efe8c"
dependencies = [
"autocfg",
"cfg-if",
"crossbeam-utils",
"autocfg 1.1.0",
"cfg-if 1.0.0",
"crossbeam-utils 0.8.8",
"lazy_static",
"memoffset",
"scopeguard",
]
[[package]]
name = "crossbeam-queue"
version = "0.1.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "7c979cd6cfe72335896575c6b5688da489e420d36a27a0b9eb0c73db574b4a4b"
dependencies = [
"crossbeam-utils 0.6.6",
]
[[package]]
name = "crossbeam-utils"
version = "0.6.6"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "04973fa96e96579258a5091af6003abde64af786b860f18622b82e026cca60e6"
dependencies = [
"cfg-if 0.1.10",
"lazy_static",
]
[[package]]
name = "crossbeam-utils"
version = "0.8.8"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "0bf124c720b7686e3c2663cf54062ab0f68a88af2fb6a030e87e30bf721fcb38"
dependencies = [
"cfg-if",
"cfg-if 1.0.0",
"lazy_static",
]
@@ -603,7 +669,7 @@ version = "4.0.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "e77a43b28d0668df09411cb0bc9a8c2adc40f9a048afe863e05fd43251e8e39c"
dependencies = [
"cfg-if",
"cfg-if 1.0.0",
"num_cpus",
]
@@ -633,7 +699,7 @@ version = "2.0.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "b98cf8ebf19c3d1b223e151f99a4f9f0690dca41414773390fc824184ac833e1"
dependencies = [
"cfg-if",
"cfg-if 1.0.0",
"dirs-sys-next",
]
@@ -672,7 +738,7 @@ version = "0.8.30"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "7896dc8abb250ffdda33912550faa54c88ec8b998dec0b2c55ab224921ce11df"
dependencies = [
"cfg-if",
"cfg-if 1.0.0",
]
[[package]]
@@ -716,6 +782,18 @@ dependencies = [
"synstructure",
]
[[package]]
name = "fallible-iterator"
version = "0.2.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "4443176a9f2c162692bd3d352d745ef9413eec5782a80d8fd6f8a1ac692a07f7"
[[package]]
name = "fallible-streaming-iterator"
version = "0.1.9"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "7360491ce676a36bf9bb3c56c1aa791658183a54d2744120f27285738d90465a"
[[package]]
name = "fastrand"
version = "1.7.0"
@@ -885,10 +963,13 @@ dependencies = [
"aws-sdk-s3",
"base64",
"bytes 1.1.0",
"bytesize",
"chrono",
"futures",
"futures-util",
"garage_api",
"garage_block",
"garage_db",
"garage_model 0.7.0",
"garage_rpc 0.7.0",
"garage_table 0.7.0",
@@ -911,9 +992,9 @@ dependencies = [
"serde_bytes",
"serde_json",
"sha2",
"sled",
"static_init",
"structopt",
"timeago",
"tokio",
"toml",
"tracing",
@@ -968,10 +1049,12 @@ dependencies = [
name = "garage_block"
version = "0.7.0"
dependencies = [
"arc-swap",
"async-trait",
"bytes 1.1.0",
"futures",
"futures-util",
"garage_db",
"garage_rpc 0.7.0",
"garage_table 0.7.0",
"garage_util 0.7.0",
@@ -981,12 +1064,26 @@ dependencies = [
"rmp-serde 0.15.5",
"serde",
"serde_bytes",
"sled",
"tokio",
"tracing",
"zstd",
]
[[package]]
name = "garage_db"
version = "0.8.0"
dependencies = [
"clap 3.1.18",
"err-derive 0.3.1",
"heed",
"hexdump",
"mktemp",
"pretty_env_logger",
"rusqlite",
"sled",
"tracing",
]
[[package]]
name = "garage_model"
version = "0.5.1"
@@ -1024,6 +1121,7 @@ dependencies = [
"futures",
"futures-util",
"garage_block",
"garage_db",
"garage_model 0.5.1",
"garage_rpc 0.7.0",
"garage_table 0.7.0",
@@ -1035,7 +1133,6 @@ dependencies = [
"rmp-serde 0.15.5",
"serde",
"serde_bytes",
"sled",
"tokio",
"tracing",
"zstd",
@@ -1130,6 +1227,7 @@ dependencies = [
"bytes 1.1.0",
"futures",
"futures-util",
"garage_db",
"garage_rpc 0.7.0",
"garage_util 0.7.0",
"hexdump",
@@ -1138,7 +1236,6 @@ dependencies = [
"rmp-serde 0.15.5",
"serde",
"serde_bytes",
"sled",
"tokio",
"tracing",
]
@@ -1173,10 +1270,12 @@ dependencies = [
name = "garage_util"
version = "0.7.0"
dependencies = [
"async-trait",
"blake2",
"chrono",
"err-derive 0.3.1",
"futures",
"garage_db",
"hex",
"http",
"hyper",
@@ -1187,7 +1286,6 @@ dependencies = [
"serde",
"serde_json",
"sha2",
"sled",
"tokio",
"toml",
"tracing",
@@ -1237,7 +1335,7 @@ version = "0.2.5"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d39cd93900197114fa1fcb7ae84ca742095eed9442088988ae74fa744e930e77"
dependencies = [
"cfg-if",
"cfg-if 1.0.0",
"libc",
"wasi 0.10.0+wasi-snapshot-preview1",
]
@@ -1288,6 +1386,18 @@ name = "hashbrown"
version = "0.11.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "ab5ef0d4909ef3724cc8cce6ccc8572c5c817592e9285f5464f8e86f8bd3726e"
dependencies = [
"ahash",
]
[[package]]
name = "hashlink"
version = "0.7.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "7249a3129cbc1ffccd74857f81464a323a152173cdb134e0fd81bc803b29facf"
dependencies = [
"hashbrown",
]
[[package]]
name = "heck"
@@ -1304,6 +1414,45 @@ version = "0.4.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "2540771e65fc8cb83cd6e8a237f70c319bd5c29f78ed1084ba5d50eeac86f7f9"
[[package]]
name = "heed"
version = "0.11.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "269c7486ed6def5d7b59a427cec3e87b4d4dd4381d01e21c8c9f2d3985688392"
dependencies = [
"bytemuck",
"byteorder",
"heed-traits",
"heed-types",
"libc",
"lmdb-rkv-sys",
"once_cell",
"page_size",
"serde",
"synchronoise",
"url",
]
[[package]]
name = "heed-traits"
version = "0.8.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "a53a94e5b2fd60417e83ffdfe136c39afacff0d4ac1d8d01cd66928ac610e1a2"
[[package]]
name = "heed-types"
version = "0.8.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9a6cf0a6952fcedc992602d5cddd1e3fff091fbe87d38636e3ec23a31f32acbd"
dependencies = [
"bincode",
"bytemuck",
"byteorder",
"heed-traits",
"serde",
"serde_json",
]
[[package]]
name = "hermit-abi"
version = "0.1.19"
@@ -1493,7 +1642,7 @@ version = "1.8.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "282a6247722caba404c065016bbfa522806e51714c34f5dfc3e4a3a46fcb4223"
dependencies = [
"autocfg",
"autocfg 1.1.0",
"hashbrown",
]
@@ -1503,7 +1652,7 @@ version = "0.1.12"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "7a5bbe824c507c5da5956355e86a746d82e0e1464f65d862cc5e71da70e94b2c"
dependencies = [
"cfg-if",
"cfg-if 1.0.0",
]
[[package]]
@@ -1515,6 +1664,16 @@ dependencies = [
"serde",
]
[[package]]
name = "isolang"
version = "1.0.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "265ef164908329e47e753c769b14cbb27434abf0c41984dca201484022f09ce5"
dependencies = [
"phf",
"phf_codegen",
]
[[package]]
name = "itertools"
version = "0.4.19"
@@ -1578,7 +1737,7 @@ dependencies = [
[[package]]
name = "k2v-client"
version = "0.1.0"
version = "0.0.1"
dependencies = [
"base64",
"clap 3.1.18",
@@ -1758,12 +1917,34 @@ dependencies = [
"walkdir",
]
[[package]]
name = "libsqlite3-sys"
version = "0.24.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "898745e570c7d0453cc1fbc4a701eb6c662ed54e8fec8b7d14be137ebeeb9d14"
dependencies = [
"cc",
"pkg-config",
"vcpkg",
]
[[package]]
name = "linked-hash-map"
version = "0.5.4"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "7fb9b38af92608140b86b693604b9ffcc5824240a484d1ecd4795bacb2fe88f3"
[[package]]
name = "lmdb-rkv-sys"
version = "0.11.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "61b9ce6b3be08acefa3003c57b7565377432a89ec24476bbe72e11d101f852fe"
dependencies = [
"cc",
"libc",
"pkg-config",
]
[[package]]
name = "lock_api"
version = "0.4.6"
@@ -1779,7 +1960,7 @@ version = "0.4.16"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "6389c490849ff5bc16be905ae24bc913a9c8892e19b2341dbc175e14c341c2b8"
dependencies = [
"cfg-if",
"cfg-if 1.0.0",
]
[[package]]
@@ -1817,7 +1998,7 @@ version = "0.6.5"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "5aa361d4faea93603064a027415f07bd8e1d5c88c9fbf68bf56a285428fd79ce"
dependencies = [
"autocfg",
"autocfg 1.1.0",
]
[[package]]
@@ -1855,6 +2036,15 @@ dependencies = [
"winapi",
]
[[package]]
name = "mktemp"
version = "0.4.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "975de676448231fcde04b9149d2543077e166b78fc29eae5aa219e7928410da2"
dependencies = [
"uuid",
]
[[package]]
name = "multer"
version = "2.0.2"
@@ -1928,7 +2118,7 @@ dependencies = [
"arc-swap",
"async-trait",
"bytes 0.6.0",
"cfg-if",
"cfg-if 1.0.0",
"err-derive 0.2.4",
"futures",
"hex",
@@ -1970,7 +2160,7 @@ version = "0.1.44"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d2cc698a63b549a70bc047073d2949cce27cd1c7b0a4a862d08a8031bc2801db"
dependencies = [
"autocfg",
"autocfg 1.1.0",
"num-traits",
]
@@ -1980,7 +2170,7 @@ version = "0.2.14"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9a64b1ec5cda2586e284722486d802acf1f7dbdc623e2bfc57e65ca1cd099290"
dependencies = [
"autocfg",
"autocfg 1.1.0",
]
[[package]]
@@ -2021,7 +2211,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "0c7ae222234c30df141154f159066c5093ff73b63204dcda7121eb082fc56a95"
dependencies = [
"bitflags",
"cfg-if",
"cfg-if 1.0.0",
"foreign-types",
"libc",
"once_cell",
@@ -2049,7 +2239,7 @@ version = "0.9.72"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "7e46109c383602735fa0a2e48dd2b7c892b048e1bf69e5c3b1d804b7d9c203cb"
dependencies = [
"autocfg",
"autocfg 1.1.0",
"cc",
"libc",
"openssl-src",
@@ -2134,6 +2324,16 @@ version = "6.0.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "029d8d0b2f198229de29dca79676f2738ff952edf3fde542eb8bf94d8c21b435"
[[package]]
name = "page_size"
version = "0.4.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "eebde548fbbf1ea81a99b128872779c437752fb99f217c45245e1a61dcd9edcd"
dependencies = [
"libc",
"winapi",
]
[[package]]
name = "parking_lot"
version = "0.11.2"
@@ -2161,7 +2361,7 @@ version = "0.8.5"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d76e8e1493bcac0d2766c42737f34458f1c8c50c0d23bcb24ea953affb273216"
dependencies = [
"cfg-if",
"cfg-if 1.0.0",
"instant",
"libc",
"redox_syscall",
@@ -2175,7 +2375,7 @@ version = "0.9.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "28141e0cc4143da2443301914478dc976a61ffdb3f043058310c70df2fed8954"
dependencies = [
"cfg-if",
"cfg-if 1.0.0",
"libc",
"redox_syscall",
"smallvec",
@@ -2209,6 +2409,44 @@ dependencies = [
"indexmap",
]
[[package]]
name = "phf"
version = "0.7.24"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "b3da44b85f8e8dfaec21adae67f95d93244b2ecf6ad2a692320598dcc8e6dd18"
dependencies = [
"phf_shared",
]
[[package]]
name = "phf_codegen"
version = "0.7.24"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "b03e85129e324ad4166b06b2c7491ae27fe3ec353af72e72cd1654c7225d517e"
dependencies = [
"phf_generator",
"phf_shared",
]
[[package]]
name = "phf_generator"
version = "0.7.24"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "09364cc93c159b8b06b1f4dd8a4398984503483891b0c26b867cf431fb132662"
dependencies = [
"phf_shared",
"rand 0.6.5",
]
[[package]]
name = "phf_shared"
version = "0.7.24"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "234f71a15de2288bcb7e3b6515828d22af7ec8598ee6d24c3b526fa0a80b67a0"
dependencies = [
"siphasher",
]
[[package]]
name = "pin-project"
version = "0.4.29"
@@ -2357,7 +2595,7 @@ version = "0.13.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "b7f64969ffd5dd8f39bd57a68ac53c163a095ed9d0fb707146da1b27025a3504"
dependencies = [
"cfg-if",
"cfg-if 1.0.0",
"fnv",
"lazy_static",
"memchr",
@@ -2463,6 +2701,25 @@ dependencies = [
"winapi",
]
[[package]]
name = "rand"
version = "0.6.5"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "6d71dacdc3c88c1fde3885a3be3fbab9f35724e6ce99467f7d9c5026132184ca"
dependencies = [
"autocfg 0.1.8",
"libc",
"rand_chacha 0.1.1",
"rand_core 0.4.2",
"rand_hc",
"rand_isaac",
"rand_jitter",
"rand_os",
"rand_pcg",
"rand_xorshift",
"winapi",
]
[[package]]
name = "rand"
version = "0.8.5"
@@ -2470,10 +2727,20 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "34af8d1a0e25924bc5b7c43c079c942339d8f0a8b57c39049bef581b46327404"
dependencies = [
"libc",
"rand_chacha",
"rand_chacha 0.3.1",
"rand_core 0.6.3",
]
[[package]]
name = "rand_chacha"
version = "0.1.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "556d3a1ca6600bfcbab7c7c91ccb085ac7fbbcd70e008a98742e7847f4f7bcef"
dependencies = [
"autocfg 0.1.8",
"rand_core 0.3.1",
]
[[package]]
name = "rand_chacha"
version = "0.3.1"
@@ -2508,6 +2775,77 @@ dependencies = [
"getrandom",
]
[[package]]
name = "rand_hc"
version = "0.1.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "7b40677c7be09ae76218dc623efbf7b18e34bced3f38883af07bb75630a21bc4"
dependencies = [
"rand_core 0.3.1",
]
[[package]]
name = "rand_isaac"
version = "0.1.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "ded997c9d5f13925be2a6fd7e66bf1872597f759fd9dd93513dd7e92e5a5ee08"
dependencies = [
"rand_core 0.3.1",
]
[[package]]
name = "rand_jitter"
version = "0.1.4"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "1166d5c91dc97b88d1decc3285bb0a99ed84b05cfd0bc2341bdf2d43fc41e39b"
dependencies = [
"libc",
"rand_core 0.4.2",
"winapi",
]
[[package]]
name = "rand_os"
version = "0.1.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "7b75f676a1e053fc562eafbb47838d67c84801e38fc1ba459e8f180deabd5071"
dependencies = [
"cloudabi",
"fuchsia-cprng",
"libc",
"rand_core 0.4.2",
"rdrand",
"winapi",
]
[[package]]
name = "rand_pcg"
version = "0.1.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "abf9b09b01790cfe0364f52bf32995ea3c39f4d2dd011eac241d2914146d0b44"
dependencies = [
"autocfg 0.1.8",
"rand_core 0.4.2",
]
[[package]]
name = "rand_xorshift"
version = "0.1.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "cbf7e9e623549b0e21f6e97cf8ecf247c1a8fd2e8a992ae265314300b2455d5c"
dependencies = [
"rand_core 0.3.1",
]
[[package]]
name = "rdrand"
version = "0.4.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "678054eb77286b51581ba43620cc911abf02758c91f93f479767aed0f90458b2"
dependencies = [
"rand_core 0.3.1",
]
[[package]]
name = "redox_syscall"
version = "0.2.11"
@@ -2679,6 +3017,21 @@ dependencies = [
"tokio",
]
[[package]]
name = "rusqlite"
version = "0.27.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "85127183a999f7db96d1a976a309eebbfb6ea3b0b400ddd8340190129de6eb7a"
dependencies = [
"bitflags",
"fallible-iterator",
"fallible-streaming-iterator",
"hashlink",
"libsqlite3-sys",
"memchr",
"smallvec",
]
[[package]]
name = "rustc_version"
version = "0.4.0"
@@ -2894,7 +3247,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "4d58a1e1bf39749807d89cf2d98ac2dfa0ff1cb3faa38fbb64dd88ac8013d800"
dependencies = [
"block-buffer",
"cfg-if",
"cfg-if 1.0.0",
"cpufeatures",
"digest",
"opaque-debug",
@@ -2915,6 +3268,12 @@ dependencies = [
"libc",
]
[[package]]
name = "siphasher"
version = "0.2.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "0b8de496cf83d4ed58b6be86c3a275b8602f6ffe98d3024a869e124147a9a3ac"
[[package]]
name = "slab"
version = "0.4.5"
@@ -2929,7 +3288,7 @@ checksum = "7f96b4737c2ce5987354855aed3797279def4ebf734436c6aa4552cf8e169935"
dependencies = [
"crc32fast",
"crossbeam-epoch",
"crossbeam-utils",
"crossbeam-utils 0.8.8",
"fs2",
"fxhash",
"libc",
@@ -3063,6 +3422,15 @@ dependencies = [
"unicode-xid",
]
[[package]]
name = "synchronoise"
version = "1.0.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d717ed0efc9d39ab3b642a096bc369a3e02a38a51c41845d7fe31bdad1d6eaeb"
dependencies = [
"crossbeam-queue",
]
[[package]]
name = "synstructure"
version = "0.12.6"
@@ -3081,7 +3449,7 @@ version = "3.3.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "5cdb1ef4eaeeaddc8fbd371e5017057064af0911902ef36b39801f67cc6d79e4"
dependencies = [
"cfg-if",
"cfg-if 1.0.0",
"fastrand",
"libc",
"redox_syscall",
@@ -3154,6 +3522,16 @@ dependencies = [
"num_threads",
]
[[package]]
name = "timeago"
version = "0.3.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "6ec32dde57efb15c035ac074118d7f32820451395f28cb0524a01d4e94983b26"
dependencies = [
"chrono",
"isolang",
]
[[package]]
name = "tinyvec"
version = "1.5.1"
@@ -3380,7 +3758,7 @@ version = "0.1.32"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "4a1bdf54a7c28a2bbf701e1d2233f6c77f473486b94bee4f9678da5a148dca7f"
dependencies = [
"cfg-if",
"cfg-if 1.0.0",
"log",
"pin-project-lite",
"tracing-attributes",
@@ -3489,6 +3867,15 @@ dependencies = [
"percent-encoding",
]
[[package]]
name = "uuid"
version = "0.8.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "bc5cf98d8186244414c848017f0e2676b3fcb46807f6668a97dfe67359a3c4b7"
dependencies = [
"getrandom",
]
[[package]]
name = "vcpkg"
version = "0.2.15"
@@ -3540,7 +3927,7 @@ version = "0.2.79"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "25f1af7423d8588a3d840681122e72e6a24ddbcb3f0ec385cac0d12d24256c06"
dependencies = [
"cfg-if",
"cfg-if 1.0.0",
"wasm-bindgen-macro",
]
+715 -191
View File
File diff suppressed because it is too large Load Diff
+1
View File
@@ -1,5 +1,6 @@
[workspace]
members = [
"src/db",
"src/util",
"src/rpc",
"src/table",
+18 -4
View File
@@ -1,13 +1,27 @@
.PHONY: doc all release shell
.PHONY: doc all release shell run1 run2 run3
all:
clear; cargo build --all-features
doc:
cd doc/book; mdbook build
release:
nix-build --arg release true
shell:
nix-shell
# ----
run1:
RUST_LOG=garage=debug ./target/debug/garage -c tmp/config1.toml server
run1rel:
RUST_LOG=garage=debug ./target/release/garage -c tmp/config1.toml server
run2:
RUST_LOG=garage=debug ./target/debug/garage -c tmp/config2.toml server
run2rel:
RUST_LOG=garage=debug ./target/release/garage -c tmp/config2.toml server
run3:
RUST_LOG=garage=debug ./target/debug/garage -c tmp/config3.toml server
run3rel:
RUST_LOG=garage=debug ./target/release/garage -c tmp/config3.toml server
+4 -3
View File
@@ -69,7 +69,7 @@ in let
we ask the user (the CI often) to pass the value to Nix.
*/
(pkgs.rustBuilder.rustLib.makeOverride {
name = "garage";
name = "garage_rpc";
overrideAttrs = drv:
/* [1] */ { hardeningDisable = [ "pie" ]; }
//
@@ -86,13 +86,14 @@ in let
It speeds up the compilation (when the feature is not required) and released crates have less dependency by default (less attack surface, disk space, etc.).
But we want to ship these additional features when we release Garage.
In the end, we chose to exclude all features from debug builds while putting (all of) them in the release builds.
Currently, the only feature of Garage is kubernetes-discovery from the garage_rpc crate.
Currently, the only such feature of Garage is kubernetes-discovery from the garage_rpc crate.
The experimental feature k2v is also enabled here in all builds.
*/
(pkgs.rustBuilder.rustLib.makeOverride {
name = "garage_rpc";
overrideArgs = old:
{
features = if release then [ "kubernetes-discovery" ] else [];
features = if release then [ "kubernetes-discovery" "k2v" ] else [ "k2v" ];
};
})
];
+55
View File
@@ -17,6 +17,61 @@ If you still want to use Borg, you can use it with `rclone mount`.
## Restic
Create your key and bucket:
```bash
garage key new my-key
garage bucket create backup
garage bucket allow backup --read --write --key my-key
```
Then register your Key ID and Secret key in your environment:
```bash
export AWS_ACCESS_KEY_ID=GKxxx
export AWS_SECRET_ACCESS_KEY=xxxx
```
Configure restic from environment too:
```bash
export RESTIC_REPOSITORY="s3:http://localhost:3900/backups"
echo "Generated password (save it safely): $(openssl rand -base64 32)"
export RESTIC_PASSWORD=xxx # copy paste your generated password here
```
Do not forget to save your password safely (in your password manager or print it). It will be needed to decrypt your backups.
Now you can use restic:
```bash
# Initialize the bucket, must be run once
restic init
# Backup your PostgreSQL database
# (We suppose your PostgreSQL daemon is stopped for all commands)
restic backup /var/lib/postgresql
# Show backup history
restic snapshots
# Backup again your PostgreSQL database, it will be faster as only changes will be uploaded
restic backup /var/lib/postgresql
# Show backup history (again)
restic snapshots
# Restore a backup
# (79766175 is the ID of the snapshot you want to restore)
mv /var/lib/postgresql /var/lib/postgresql.broken
restic restore 79766175 --target /var/lib/postgresql
```
Restic has way more features than the ones presented here.
You can discover all of them by accessing its documentation from the link below.
*External links:* [Restic Documentation > Amazon S3](https://restic.readthedocs.io/en/stable/030_preparing_a_new_repo.html#amazon-s3)
## Duplicity
+1 -1
View File
@@ -3,7 +3,7 @@ title = "Websites (Hugo, Jekyll, Publii...)"
weight = 10
+++
Garage is also suitable to host static websites.
Garage is also suitable [to host static websites](@/documentation/cookbook/exposing-websites.md).
While they can be deployed with traditional CLI tools, some static website generators have integrated options to ease your workflow.
| Name | Status | Note |
+140 -3
View File
@@ -100,7 +100,7 @@ server {
}
```
## Exposing the web endpoint
### Exposing the web endpoint
To better understand the logic involved, you can refer to the [Exposing buckets as websites](/cookbook/exposing_websites.html) section.
Otherwise, the configuration is very similar to the S3 endpoint.
@@ -140,6 +140,143 @@ server {
@TODO
## Traefik
## Traefik v2
@TODO
We will see in this part how to set up a reverse proxy with [Traefik](https://docs.traefik.io/).
Here is [a basic configuration file](https://doc.traefik.io/traefik/https/acme/#configuration-examples):
```toml
[entryPoints]
[entryPoints.web]
address = ":80"
[entryPoints.websecure]
address = ":443"
[certificatesResolvers.myresolver.acme]
email = "your-email@example.com"
storage = "acme.json"
[certificatesResolvers.myresolver.acme.httpChallenge]
# used during the challenge
entryPoint = "web"
```
### Add Garage service
To add Garage on Traefik you should declare a new service using its IP address (or hostname) and port:
```toml
[http.services]
[http.services.my_garage_service.loadBalancer]
[[http.services.my_garage_service.loadBalancer.servers]]
url = "http://xxx.xxx.xxx.xxx"
port = 3900
```
It's possible to declare multiple Garage servers as back-ends:
```toml
[http.services]
[[http.services.my_garage_service.loadBalancer.servers]]
url = "http://xxx.xxx.xxx.xxx"
port = 3900
[[http.services.my_garage_service.loadBalancer.servers]]
url = "http://yyy.yyy.yyy.yyy"
port = 3900
[[http.services.my_garage_service.loadBalancer.servers]]
url = "http://zzz.zzz.zzz.zzz"
port = 3900
```
Traefik can remove unhealthy servers automatically with [a health check configuration](https://doc.traefik.io/traefik/routing/services/#health-check):
```
[http.services]
[http.services.my_garage_service.loadBalancer]
[http.services.my_garage_service.loadBalancer.healthCheck]
path = "/"
interval = "60s"
timeout = "5s"
```
### Adding a website
To add a new website, add the following declaration to your Traefik configuration file:
```toml
[http.routers]
[http.routers.my_website]
rule = "Host(`yoururl.example.org`)"
service = "my_garage_service"
entryPoints = ["web"]
```
Enable HTTPS access to your website with the following configuration section ([documentation](https://doc.traefik.io/traefik/https/overview/)):
```toml
...
entryPoints = ["websecure"]
[http.routers.my_website.tls]
certResolver = "myresolver"
...
```
### Adding gzip compression
Add the following configuration section [to compress response](https://doc.traefik.io/traefik/middlewares/http/compress/) using [gzip](https://developer.mozilla.org/en-US/docs/Glossary/GZip_compression) before sending them to the client:
```toml
[http.routers]
[http.routers.my_website]
...
middlewares = ["gzip_compress"]
...
[http.middlewares]
[http.middlewares.gzip_compress.compress]
```
### Add caching response
Traefik's caching middleware is only available on [entreprise version](https://doc.traefik.io/traefik-enterprise/middlewares/http-cache/), however the freely-available [Souin plugin](https://github.com/darkweak/souin#tr%C3%A6fik-container) can also do the job. (section to be completed)
### Complete example
```toml
[entryPoints]
[entryPoints.web]
address = ":80"
[entryPoints.websecure]
address = ":443"
[certificatesResolvers.myresolver.acme]
email = "your-email@example.com"
storage = "acme.json"
[certificatesResolvers.myresolver.acme.httpChallenge]
# used during the challenge
entryPoint = "web"
[http.routers]
[http.routers.my_website]
rule = "Host(`yoururl.example.org`)"
service = "my_garage_service"
middlewares = ["gzip_compress"]
entryPoints = ["websecure"]
[http.services]
[http.services.my_garage_service.loadBalancer]
[http.services.my_garage_service.loadBalancer.healthCheck]
path = "/"
interval = "60s"
timeout = "5s"
[[http.services.my_garage_service.loadBalancer.servers]]
url = "http://xxx.xxx.xxx.xxx"
[[http.services.my_garage_service.loadBalancer.servers]]
url = "http://yyy.yyy.yyy.yyy"
[[http.services.my_garage_service.loadBalancer.servers]]
url = "http://zzz.zzz.zzz.zzz"
[http.middlewares]
[http.middlewares.gzip_compress.compress]
```
-10
View File
@@ -249,16 +249,6 @@ mc alias set \
--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.
@@ -1,18 +1,41 @@
# Specification of Garage's administration API
+++
title = "Administration API"
weight = 16
+++
The Garage administration API is accessible through a dedicated server whose
listen address is specified in the `[admin]` section of the configuration
file (see [configuration file
reference](@/documentation/reference-manual/configuration.md))
**WARNING.** At this point, there is no comittement to stability of the APIs described in this document.
We will bump the version numbers prefixed to each API endpoint at each time the syntax
or semantics change, meaning that code that relies on these endpoint will break
when changes are introduced.
The Garage administration API was introduced in version 0.7.2, this document
does not apply to older versions of Garage.
## Access control
The admin API uses two different tokens for acces control, that are specified in the config file's `[admin]` section:
- `metrics_token`: the token for accessing the Metrics endpoint (if this token is not set in the config file, the Metrics endpoint can be accessed without access control);
- `admin_token`: the token for accessing all of the other administration endpoints (if this token is not set in the config file, access to these endpoints is disabled entirely).
- `metrics_token`: the token for accessing the Metrics endpoint (if this token
is not set in the config file, the Metrics endpoint can be accessed without
access control);
- `admin_token`: the token for accessing all of the other administration
endpoints (if this token is not set in the config file, access to these
endpoints is disabled entirely).
These tokens are used as simple HTTP bearer tokens. In other words, to
authenticate access to an admin API endpoint, add the following HTTP header
to your request:
```
Authorization: Bearer <token>
```
## Administration API endpoints
@@ -111,8 +134,8 @@ Example request body:
```json
[
"ec79480e0ce52ae26fd00c9da684e4fa56658d9c64cdcecb094e936de0bfe71f@10.0.0.11:3901",
"4a6ae5a1d0d33bf895f5bb4f0a418b7dc94c47c0dd2eb108d1158f3c8f60b0ff@10.0.0.12:3901"
"ec79480e0ce52ae26fd00c9da684e4fa56658d9c64cdcecb094e936de0bfe71f@10.0.0.11:3901",
"4a6ae5a1d0d33bf895f5bb4f0a418b7dc94c47c0dd2eb108d1158f3c8f60b0ff@10.0.0.12:3901"
]
```
@@ -122,14 +145,14 @@ Example response:
```json
[
{
"success": true,
"error": null
},
{
"success": false,
"error": "Handshake error"
}
{
"success": true,
"error": null
},
{
"success": false,
"error": "Handshake error"
}
]
```
@@ -278,7 +301,7 @@ Request body format:
```json
{
"name": "NameOfMyKey"
"name": "NameOfMyKey"
}
```
@@ -290,9 +313,9 @@ Request body format:
```json
{
"accessKeyId": "GK31c2f218a2e44f485b94239e",
"secretAccessKey": "b892c0665f0ada8a4755dae98baa3b133590e11dae3bcc1f9d769d67f16c3835",
"name": "NameOfMyKey"
"accessKeyId": "GK31c2f218a2e44f485b94239e",
"secretAccessKey": "b892c0665f0ada8a4755dae98baa3b133590e11dae3bcc1f9d769d67f16c3835",
"name": "NameOfMyKey"
}
```
@@ -380,11 +403,11 @@ Request body format:
```json
{
"name": "NameOfMyKey",
"allow": {
"createBucket": true,
},
"deny": {}
"name": "NameOfMyKey",
"allow": {
"createBucket": true,
},
"deny": {}
}
```
@@ -450,24 +473,31 @@ Example response:
```json
{
"id": "e6a14cd6a27f48684579ec6b381c078ab11697e6bc8513b72b2f5307e25fff9b",
"globalAliases": [
"alex"
],
"keys": [
{
"accessKeyId": "GK31c2f218a2e44f485b94239e",
"name": "alex",
"permissions": {
"read": true,
"write": true,
"owner": true
},
"bucketLocalAliases": [
"test"
]
}
]
"id": "afa8f0a22b40b1247ccd0affb869b0af5cff980924a20e4b5e0720a44deb8d39",
"globalAliases": [],
"websiteAccess": false,
"websiteConfig": null,
"keys": [
{
"accessKeyId": "GK31c2f218a2e44f485b94239e",
"name": "Imported key",
"permissions": {
"read": true,
"write": true,
"owner": true
},
"bucketLocalAliases": [
"debug"
]
}
],
"objects": 14827,
"bytes": 13189855625,
"unfinshedUploads": 0,
"quotas": {
"maxSize": null,
"maxObjects": null
}
}
```
@@ -479,7 +509,7 @@ Request body format:
```json
{
"globalAlias": "NameOfMyBucket"
"globalAlias": "NameOfMyBucket"
}
```
@@ -487,15 +517,15 @@ OR
```json
{
"localAlias": {
"accessKeyId": "GK31c2f218a2e44f485b94239e",
"alias": "NameOfMyBucket",
"allow": {
"read": true,
"write": true,
"owner": false
}
}
"localAlias": {
"accessKeyId": "GK31c2f218a2e44f485b94239e",
"alias": "NameOfMyBucket",
"allow": {
"read": true,
"write": true,
"owner": false
}
}
}
```
@@ -517,26 +547,37 @@ Deletes a storage bucket. A bucket cannot be deleted if it is not empty.
Warning: this will delete all aliases associated with the bucket!
#### PutBucketWebsite `PUT /v0/bucket/website?id=<bucket id>`
#### UpdateBucket `PUT /v0/bucket?id=<bucket id>`
Sets the website configuration for a bucket (this also enables website access for this bucket).
Updates configuration of the given bucket.
Request body format:
```json
{
"indexDocument": "index.html",
"errorDocument": "404.html"
"websiteAccess": {
"enabled": true,
"indexDocument": "index.html",
"errorDocument": "404.html"
},
"quotas": {
"maxSize": 19029801,
"maxObjects": null,
}
}
```
The field `errorDocument` is optional, if no error document is set a generic error message is displayed when errors happen.
All fields (`websiteAccess` and `quotas`) are optionnal.
If they are present, the corresponding modifications are applied to the bucket, otherwise nothing is changed.
In `websiteAccess`: if `enabled` is `true`, `indexDocument` must be specified.
The field `errorDocument` is optional, if no error document is set a generic
error message is displayed when errors happen. Conversely, if `enabled` is
`false`, neither `indexDocument` nor `errorDocument` must be specified.
#### DeleteBucketWebsite `DELETE /v0/bucket/website?id=<bucket id>`
Deletes the website configuration for a bucket (disables website access for this bucket).
In `quotas`: new values of `maxSize` and `maxObjects` must both be specified, or set to `null`
to remove the quotas. An absent value will be considered the same as a `null`. It is not possible
to change only one of the two quotas.
### Operations on permissions for keys on buckets
@@ -548,13 +589,13 @@ Request body format:
```json
{
"bucketId": "e6a14cd6a27f48684579ec6b381c078ab11697e6bc8513b72b2f5307e25fff9b",
"accessKeyId": "GK31c2f218a2e44f485b94239e",
"permissions": {
"read": true,
"write": true,
"owner": true
},
"bucketId": "e6a14cd6a27f48684579ec6b381c078ab11697e6bc8513b72b2f5307e25fff9b",
"accessKeyId": "GK31c2f218a2e44f485b94239e",
"permissions": {
"read": true,
"write": true,
"owner": true
},
}
```
@@ -569,13 +610,13 @@ Request body format:
```json
{
"bucketId": "e6a14cd6a27f48684579ec6b381c078ab11697e6bc8513b72b2f5307e25fff9b",
"accessKeyId": "GK31c2f218a2e44f485b94239e",
"permissions": {
"read": false,
"write": false,
"owner": true
},
"bucketId": "e6a14cd6a27f48684579ec6b381c078ab11697e6bc8513b72b2f5307e25fff9b",
"accessKeyId": "GK31c2f218a2e44f485b94239e",
"permissions": {
"read": false,
"write": false,
"owner": true
},
}
```
+32 -4
View File
@@ -10,6 +10,7 @@ metadata_dir = "/var/lib/garage/meta"
data_dir = "/var/lib/garage/data"
block_size = 1048576
block_manager_background_tranquility = 2
replication_mode = "3"
@@ -47,6 +48,8 @@ root_domain = ".web.garage"
[admin]
api_bind_addr = "0.0.0.0:3903"
metrics_token = "cacce0b2de4bc2d9f5b5fdff551e01ac1496055aed248202d415398987e35f81"
admin_token = "ae8cb40ea7368bbdbb6430af11cca7da833d3458a5f52086f4e805a570fb5c2a"
trace_sink = "http://localhost:4317"
```
@@ -84,6 +87,17 @@ 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.
### `block_manager_background_tranquility`
This parameter tunes the activity of the background worker responsible for
resyncing data blocks between nodes. The higher the tranquility value is set,
the more the background worker will wait between iterations, meaning the load
on the system (including network usage between nodes) will be reduced. The
minimal value for this parameter is `0`, where the background worker will
allways work at maximal throughput to resynchronize blocks. The default value
is `2`, where the background worker will try to spend at most 1/3 of its time
working, and 2/3 sleeping in order to reduce system load.
### `replication_mode`
Garage supports the following replication modes:
@@ -326,10 +340,24 @@ Garage has a few administration capabilities, in particular to allow remote moni
### `api_bind_addr`
If specified, Garage will bind an HTTP server to this port and address, on
which it will listen to requests for administration features. Currently,
this endpoint only exposes Garage metrics in the Prometheus format at
`/metrics`. This endpoint is not authenticated. In the future, bucket and
access key management might be possible by REST calls to this endpoint.
which it will listen to requests for administration features.
See [administration API reference](@/documentation/reference-manual/admin-api.md) to learn more about these features.
### `metrics_token` (since version 0.7.2)
The token for accessing the Metrics endpoint. If this token is not set in
the config file, the Metrics endpoint can be accessed without access
control.
You can use any random string for this value. We recommend generating a random token with `openssl rand -hex 32`.
### `admin_token` (since version 0.7.2)
The token for accessing all of the other administration endpoints. If this
token is not set in the config file, access to these endpoints is disabled
entirely.
You can use any random string for this value. We recommend generating a random token with `openssl rand -hex 32`.
### `trace_sink`
+58
View File
@@ -0,0 +1,58 @@
+++
title = "K2V"
weight = 30
+++
Starting with version 0.7.2, Garage introduces an optionnal feature, K2V,
which is an alternative storage API designed to help efficiently store
many small values in buckets (in opposition to S3 which is more designed
to store large blobs).
K2V is currently disabled at compile time in all builds, as the
specification is still subject to changes. To build a Garage version with
K2V, the Cargo feature flag `k2v` must be activated. Special builds with
the `k2v` feature flag enabled can be obtained from our download page under
"Extra builds": such builds can be identified easily as their tag name ends
with `-k2v` (example: `v0.7.2-k2v`).
The specification of the K2V API can be found
[here](https://git.deuxfleurs.fr/Deuxfleurs/garage/src/branch/k2v/doc/drafts/k2v-spec.md).
This document also includes a high-level overview of K2V's design.
The K2V API uses AWSv4 signatures for authentification, same as the S3 API.
The AWS region used for signature calculation is always the same as the one
defined for the S3 API in the config file.
## Enabling and using K2V
To enable K2V, download and run a build that has the `k2v` feature flag
enabled, or produce one yourself. Then, add the following section to your
configuration file:
```toml
[k2v_api]
api_bind_addr = "<ip>:<port>"
```
Please select a port number that is not already in use by another API
endpoint (S3 api, admin API) or by the RPC server.
We provide an early-stage K2V client library for Rust which can be imported by adding the following to your `Cargo.toml` file:
```toml
k2v-client = { git = "https://git.deuxfleurs.fr/Deuxfleurs/garage.git" }
```
There is also a simple CLI utility which can be built from source in the
following way:
```sh
git clone https://git.deuxfleurs.fr/Deuxfleurs/garage.git
cd garage/src/k2v-client
cargo build --features cli --bin k2v-cli
```
The CLI utility is self-documented, run `k2v-cli --help` to learn how to use
it. There is also a short README.md in the `src/k2v-client` folder with some
instructions.
+77 -50
View File
@@ -3,51 +3,77 @@ title = "S3 Compatibility status"
weight = 20
+++
## Endpoint implementation
## DISCLAIMER
All APIs that are missing on Garage will return a 501 Not Implemented.
Some `x-amz-` headers are not implemented.
**The compatibility list for other platforms is given only for informational
purposes and based on available documentation.** They are sometimes completed,
in a best effort approach, with the source code and inputs from maintainers
when documentation is lacking. We are not proactively monitoring new versions
of each software: check the modification history to know when the page has been
updated for the last time. Some entries will be inexact or outdated. For any
serious decision, you must make your own tests.
**The official documentation of each project can be accessed by clicking on the
project name in the column header.**
*The compatibility list for other platforms is given only for information purposes and based on available documentation. Some entries might be inexact. Feel free to open a PR to fix this table. Minio is missing because they do not provide a public S3 compatibility list.*
Feel free to open a PR to suggest fixes this table. Minio is missing because they do not provide a public S3 compatibility list.
### Features
## Update history
- 2022-02-07 - First version of this page
- 2022-05-25 - Many Ceph S3 endpoints are not documented but implemented. Following a notification from the Ceph community, we added them.
## High-level features
| Feature | Garage | [Openstack Swift](https://docs.openstack.org/swift/latest/s3_compat.html) | [Ceph Object Gateway](https://docs.ceph.com/en/latest/radosgw/s3/) | [Riak CS](https://docs.riak.com/riak/cs/2.1.1/references/apis/storage/s3/index.html) | [OpenIO](https://docs.openio.io/latest/source/arch-design/s3_compliancy.html) |
|------------------------------|----------------------------------|-----------------|---------------|---------|-----|
| [signature v2](https://docs.aws.amazon.com/general/latest/gr/signature-version-2.html) (deprecated) | ❌ Missing | ✅ | | ✅ | ✅ |
| [signature v2](https://docs.aws.amazon.com/general/latest/gr/signature-version-2.html) (deprecated) | ❌ Missing | ✅ | | ✅ | ✅ |
| [signature v4](https://docs.aws.amazon.com/AmazonS3/latest/API/sig-v4-authenticating-requests.html) | ✅ Implemented | ✅ | ✅ | ❌ | ✅ |
| [URL path-style](https://docs.aws.amazon.com/AmazonS3/latest/userguide/VirtualHosting.html#path-style-access) (eg. `host.tld/bucket/key`) | ✅ Implemented | ✅ | ✅ | ❓| ✅ |
| [URL vhost-style](https://docs.aws.amazon.com/AmazonS3/latest/userguide/VirtualHosting.html#virtual-hosted-style-access) URL (eg. `bucket.host.tld/key`) | ✅ Implemented | ❌| ✅| ✅ | ✅ |
| [Presigned URLs](https://docs.aws.amazon.com/AmazonS3/latest/userguide/ShareObjectPreSignedURL.html) | ✅ Implemented | ❌| ✅ | ✅ | ✅(❓) |
*Note:* OpenIO does not says if it supports presigned URLs. Because it is part of signature v4 and they claim they support it without additional precisions, we suppose that OpenIO supports presigned URLs.
*Note:* OpenIO does not says if it supports presigned URLs. Because it is part
of signature v4 and they claim they support it without additional precisions,
we suppose that OpenIO supports presigned URLs.
## Endpoint implementation
All endpoints that are missing on Garage will return a 501 Not Implemented.
Some `x-amz-` headers are not implemented.
### Core endoints
| Endpoint | Garage | [Openstack Swift](https://docs.openstack.org/swift/latest/s3_compat.html) | [Ceph Object Gateway](https://docs.ceph.com/en/latest/radosgw/s3/) | [Riak CS](https://docs.riak.com/riak/cs/2.1.1/references/apis/storage/s3/index.html) | [OpenIO](https://docs.openio.io/latest/source/arch-design/s3_compliancy.html) |
| Endpoint | Garage | [Openstack Swift](https://docs.openstack.org/swift/latest/s3_compat.html) | [Ceph Object Gateway](https://docs.ceph.com/en/latest/radosgw/s3/) | [Riak CS](https://docs.riak.com/riak/cs/2.1.1/references/apis/storage/s3/index.html) | [OpenIO](https://docs.openio.io/latest/source/arch-design/s3_compliancy.html) |
|------------------------------|----------------------------------|-----------------|---------------|---------|-----|
| [CreateBucket](https://docs.aws.amazon.com/AmazonS3/latest/API/API_CreateBucket.html) | ✅ Implemented | ✅ | ✅ | ✅ | ✅ |
| [DeleteBucket](https://docs.aws.amazon.com/AmazonS3/latest/API/API_DeleteBucket.html) | ✅ Implemented | ✅ | ✅ | ✅ | ✅ |
| [GetBucketLocation](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetBucketLocation.html) | ✅ Implemented | ✅ | ✅ | ❌ | ✅ |
| [HeadBucket](https://docs.aws.amazon.com/AmazonS3/latest/API/API_HeadBucket.html) | ✅ Implemented | ✅ | ✅ | ✅ | ✅ |
| [ListBuckets](https://docs.aws.amazon.com/AmazonS3/latest/API/API_ListBuckets.html) | ✅ Implemented | ❌| ✅ | ✅ | ✅ |
| [HeadObject](https://docs.aws.amazon.com/AmazonS3/latest/API/API_HeadObject.html) | ✅ Implemented | ✅ | ✅ | ✅ | ✅ |
| [CopyObject](https://docs.aws.amazon.com/AmazonS3/latest/API/API_CopyObject.html) | ✅ Implemented | ✅ | ✅ | ✅ | ✅ |
| [HeadObject](https://docs.aws.amazon.com/AmazonS3/latest/API/API_HeadObject.html) | ✅ Implemented | ✅ | ✅ | ✅ | ✅ |
| [CopyObject](https://docs.aws.amazon.com/AmazonS3/latest/API/API_CopyObject.html) | ✅ Implemented | ✅ | ✅ | ✅ | ✅ |
| [DeleteObject](https://docs.aws.amazon.com/AmazonS3/latest/API/API_DeleteObject.html) | ✅ Implemented | ✅ | ✅ | ✅ | ✅ |
| [DeleteObjects](https://docs.aws.amazon.com/AmazonS3/latest/API/API_DeleteObjects.html) | ✅ Implemented | ✅ | ✅ | ✅ | ✅ |
| [GetObject](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetObject.html) | ✅ Implemented | ✅ | ✅ | ✅ | ✅ |
| [ListObjects](https://docs.aws.amazon.com/AmazonS3/latest/API/API_ListObjects.html) | ✅ Implemented (see details below) | ✅ | ✅ | ✅ | ❌|
| [ListObjectsV2](https://docs.aws.amazon.com/AmazonS3/latest/API/API_ListObjectsV2.html) | ✅ Implemented | ❌| | ❌| ✅ |
| [PostObject](https://docs.aws.amazon.com/AmazonS3/latest/API/RESTObjectPOST.html) (compatibility API) | ❌ Missing | ❌| ✅ | ❌| ❌|
| [ListObjectsV2](https://docs.aws.amazon.com/AmazonS3/latest/API/API_ListObjectsV2.html) | ✅ Implemented | ❌| | ❌| ✅ |
| [PostObject](https://docs.aws.amazon.com/AmazonS3/latest/API/RESTObjectPOST.html) | ✅ Implemented | ❌| ✅ | ❌| ❌|
| [PutObject](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutObject.html) | ✅ 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.
**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.
*Note: Ceph API documentation is incomplete and miss at least HeadBucket and UploadPartCopy, but these endpoints are documented in [Red Hat Ceph Storage - Chapter 2. Ceph Object Gateway and the S3 API](https://access.redhat.com/documentation/en-us/red_hat_ceph_storage/4/html/developer_guide/ceph-object-gateway-and-the-s3-api)*
*Note: Ceph API documentation is incomplete and lacks at least HeadBucket and UploadPartCopy,
but these endpoints are documented in [Red Hat Ceph Storage - Chapter 2. Ceph Object Gateway and the S3 API](https://access.redhat.com/documentation/en-us/red_hat_ceph_storage/4/html/developer_guide/ceph-object-gateway-and-the-s3-api)*
### Multipart Upload endpoints
| Endpoint | Garage | [Openstack Swift](https://docs.openstack.org/swift/latest/s3_compat.html) | [Ceph Object Gateway](https://docs.ceph.com/en/latest/radosgw/s3/) | [Riak CS](https://docs.riak.com/riak/cs/2.1.1/references/apis/storage/s3/index.html) | [OpenIO](https://docs.openio.io/latest/source/arch-design/s3_compliancy.html) |
| Endpoint | Garage | [Openstack Swift](https://docs.openstack.org/swift/latest/s3_compat.html) | [Ceph Object Gateway](https://docs.ceph.com/en/latest/radosgw/s3/) | [Riak CS](https://docs.riak.com/riak/cs/2.1.1/references/apis/storage/s3/index.html) | [OpenIO](https://docs.openio.io/latest/source/arch-design/s3_compliancy.html) |
|------------------------------|----------------------------------|-----------------|---------------|---------|-----|
| [AbortMultipartUpload](https://docs.aws.amazon.com/AmazonS3/latest/API/API_AbortMultipartUpload.html) | ✅ Implemented | ✅ | ✅ | ✅ | ✅ |
| [CompleteMultipartUpload](https://docs.aws.amazon.com/AmazonS3/latest/API/API_CompleteMultipartUpload.html) | ✅ Implemented (see details below) | ✅ | ✅ | ✅ | ✅ |
@@ -62,18 +88,18 @@ For more information, please refer to our [issue tracker](https://git.deuxfleurs
### Website endpoints
| Endpoint | Garage | [Openstack Swift](https://docs.openstack.org/swift/latest/s3_compat.html) | [Ceph Object Gateway](https://docs.ceph.com/en/latest/radosgw/s3/) | [Riak CS](https://docs.riak.com/riak/cs/2.1.1/references/apis/storage/s3/index.html) | [OpenIO](https://docs.openio.io/latest/source/arch-design/s3_compliancy.html) |
| Endpoint | Garage | [Openstack Swift](https://docs.openstack.org/swift/latest/s3_compat.html) | [Ceph Object Gateway](https://docs.ceph.com/en/latest/radosgw/s3/) | [Riak CS](https://docs.riak.com/riak/cs/2.1.1/references/apis/storage/s3/index.html) | [OpenIO](https://docs.openio.io/latest/source/arch-design/s3_compliancy.html) |
|------------------------------|----------------------------------|-----------------|---------------|---------|-----|
| [DeleteBucketWebsite](https://docs.aws.amazon.com/AmazonS3/latest/API/API_DeleteBucketWebsite.html) | ✅ Implemented | ❌| ❌| ❌| ❌|
| [GetBucketWebsite](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetBucketWebsite.html) | ✅ Implemented | ❌ | ❌| ❌| ❌|
| [PutBucketWebsite](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutBucketWebsite.html) | ⚠ Partially implemented (see below)| ❌| ❌| ❌| ❌|
| [DeleteBucketCors](https://docs.aws.amazon.com/AmazonS3/latest/API/API_DeleteBucketCors.html) | ✅ Implemented | ❌| | ❌| ✅ |
| [GetBucketCors](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetBucketCors.html) | ✅ Implemented | ❌ | | ❌| ✅ |
| [PutBucketCors](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutBucketCors.html) | ✅ Implemented | ❌| | ❌| ✅ |
| [DeleteBucketCors](https://docs.aws.amazon.com/AmazonS3/latest/API/API_DeleteBucketCors.html) | ✅ Implemented | ❌| | ❌| ✅ |
| [GetBucketCors](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetBucketCors.html) | ✅ Implemented | ❌ | | ❌| ✅ |
| [PutBucketCors](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutBucketCors.html) | ✅ Implemented | ❌| | ❌| ✅ |
**PutBucketWebsite:** Implemented, but only stores the index document suffix and the error document path. Redirects are not supported.
*Note: Ceph radosgw has some support for static websites but it is different from Amazon one plus it does not implement its configuration endpoints.*
*Note: Ceph radosgw has some support for static websites but it is different from the Amazon one. It also does not implement its configuration endpoints.*
### ACL, Policies endpoints
@@ -81,29 +107,29 @@ Amazon has 2 access control mechanisms in S3: ACL (legacy) and policies (new one
Garage implements none of them, and has its own system instead, built around a per-access-key-per-bucket logic.
See Garage CLI reference manual to learn how to use Garage's permission system.
| Endpoint | Garage | [Openstack Swift](https://docs.openstack.org/swift/latest/s3_compat.html) | [Ceph Object Gateway](https://docs.ceph.com/en/latest/radosgw/s3/) | [Riak CS](https://docs.riak.com/riak/cs/2.1.1/references/apis/storage/s3/index.html) | [OpenIO](https://docs.openio.io/latest/source/arch-design/s3_compliancy.html) |
| Endpoint | Garage | [Openstack Swift](https://docs.openstack.org/swift/latest/s3_compat.html) | [Ceph Object Gateway](https://docs.ceph.com/en/latest/radosgw/s3/) | [Riak CS](https://docs.riak.com/riak/cs/2.1.1/references/apis/storage/s3/index.html) | [OpenIO](https://docs.openio.io/latest/source/arch-design/s3_compliancy.html) |
|------------------------------|----------------------------------|-----------------|---------------|---------|-----|
| [DeleteBucketPolicy](https://docs.aws.amazon.com/AmazonS3/latest/API/API_DeleteBucketPolicy.html) | ❌ Missing | ❌| | ✅ | ❌|
| [GetBucketPolicy](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetBucketPolicy.html) | ❌ Missing | ❌| | ⚠ | ❌|
| [GetBucketPolicyStatus](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetBucketPolicyStatus.html) | ❌ Missing | ❌| | ❌| ❌|
| [PutBucketPolicy](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutBucketPolicy.html) | ❌ Missing | ❌| | ⚠ | ❌|
| [DeleteBucketPolicy](https://docs.aws.amazon.com/AmazonS3/latest/API/API_DeleteBucketPolicy.html) | ❌ Missing | ❌| | ✅ | ❌|
| [GetBucketPolicy](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetBucketPolicy.html) | ❌ Missing | ❌| | ⚠ | ❌|
| [GetBucketPolicyStatus](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetBucketPolicyStatus.html) | ❌ Missing | ❌| | ❌| ❌|
| [PutBucketPolicy](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutBucketPolicy.html) | ❌ Missing | ❌| | ⚠ | ❌|
| [GetBucketAcl](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetBucketAcl.html) | ❌ Missing | ✅ | ✅ | ✅ | ✅ |
| [PutBucketAcl](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutBucketAcl.html) | ❌ Missing | ✅ | ✅ | ✅ | ✅ |
| [GetObjectAcl](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetObjectAcl.html) | ❌ Missing | ✅ | ✅ | ✅ | ✅ |
| [PutObjectAcl](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutObjectAcl.html) | ❌ Missing | ✅ | ✅ | ✅ | ✅ |
*Notes:* Ceph claims that it supports bucket policies but does not implement any Policy endpoints. They probably refer to their own permission system. Riak CS only supports a subset of the policy configuration.
*Notes:* Riak CS only supports a subset of the policy configuration.
### Versioning, Lifecycle endpoints
Garage does not support (yet) object versioning.
Garage does not (yet) support object versioning.
If you need this feature, please [share your use case in our dedicated issue](https://git.deuxfleurs.fr/Deuxfleurs/garage/issues/166).
| Endpoint | Garage | [Openstack Swift](https://docs.openstack.org/swift/latest/s3_compat.html) | [Ceph Object Gateway](https://docs.ceph.com/en/latest/radosgw/s3/) | [Riak CS](https://docs.riak.com/riak/cs/2.1.1/references/apis/storage/s3/index.html) | [OpenIO](https://docs.openio.io/latest/source/arch-design/s3_compliancy.html) |
| Endpoint | Garage | [Openstack Swift](https://docs.openstack.org/swift/latest/s3_compat.html) | [Ceph Object Gateway](https://docs.ceph.com/en/latest/radosgw/s3/) | [Riak CS](https://docs.riak.com/riak/cs/2.1.1/references/apis/storage/s3/index.html) | [OpenIO](https://docs.openio.io/latest/source/arch-design/s3_compliancy.html) |
|------------------------------|----------------------------------|-----------------|---------------|---------|-----|
| [DeleteBucketLifecycle](https://docs.aws.amazon.com/AmazonS3/latest/API/API_DeleteBucketLifecycle.html) | ❌ Missing | ❌| ✅| ❌| ✅|
| [GetBucketLifecycleConfiguration](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetBucketLifecycleConfiguration.html) | ❌ Missing | ❌| | ❌| ✅|
| [PutBucketLifecycleConfiguration](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutBucketLifecycleConfiguration.html) | ❌ Missing | ❌| | ❌| ✅|
| [GetBucketLifecycleConfiguration](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetBucketLifecycleConfiguration.html) | ❌ Missing | ❌| | ❌| ✅|
| [PutBucketLifecycleConfiguration](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutBucketLifecycleConfiguration.html) | ❌ Missing | ❌| | ❌| ✅|
| [GetBucketVersioning](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetBucketVersioning.html) | ❌ Stub (see below) | ✅| ✅ | ❌| ✅|
| [ListObjectVersions](https://docs.aws.amazon.com/AmazonS3/latest/API/API_ListObjectVersions.html) | ❌ Missing | ❌| ✅ | ❌| ✅|
| [PutBucketVersioning](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutBucketVersioning.html) | ❌ Missing | ❌| ✅| ❌| ✅|
@@ -111,64 +137,65 @@ If you need this feature, please [share your use case in our dedicated issue](ht
**GetBucketVersioning:** Stub implementation (Garage does not yet support versionning so this always returns "versionning not enabled").
*Note: Ceph only supports `Expiration`, `NoncurrentVersionExpiration` and `AbortIncompleteMultipartUpload` on its Lifecycle endpoints.*
### Replication endpoints
Please open an issue if you have a use case for replication.
| Endpoint | Garage | [Openstack Swift](https://docs.openstack.org/swift/latest/s3_compat.html) | [Ceph Object Gateway](https://docs.ceph.com/en/latest/radosgw/s3/) | [Riak CS](https://docs.riak.com/riak/cs/2.1.1/references/apis/storage/s3/index.html) | [OpenIO](https://docs.openio.io/latest/source/arch-design/s3_compliancy.html) |
| Endpoint | Garage | [Openstack Swift](https://docs.openstack.org/swift/latest/s3_compat.html) | [Ceph Object Gateway](https://docs.ceph.com/en/latest/radosgw/s3/) | [Riak CS](https://docs.riak.com/riak/cs/2.1.1/references/apis/storage/s3/index.html) | [OpenIO](https://docs.openio.io/latest/source/arch-design/s3_compliancy.html) |
|------------------------------|----------------------------------|-----------------|---------------|---------|-----|
| [DeleteBucketReplication](https://docs.aws.amazon.com/AmazonS3/latest/API/API_DeleteBucketReplication.html) | ❌ Missing | ❌| ✅ | ❌| ❌|
| [GetBucketReplication](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetBucketReplication.html) | ❌ Missing | ❌| ✅ | ❌| ❌|
| [PutBucketReplication](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutBucketReplication.html) | ❌ Missing | ❌| ⚠ | ❌| ❌|
*Note: Ceph documentation briefly says that Ceph supports [replication though the S3 API](https://docs.ceph.com/en/latest/radosgw/multisite-sync-policy/#s3-replication-api) but with some limitations. Additionaly, replication endpoints are not documented in the S3 compatibility page so I don't know what kind of support we can expect.*
*Note: Ceph documentation briefly says that Ceph supports
[replication through the S3 API](https://docs.ceph.com/en/latest/radosgw/multisite-sync-policy/#s3-replication-api)
but with some limitations.
Additionaly, replication endpoints are not documented in the S3 compatibility page so I don't know what kind of support we can expect.*
### Locking objects
Amazon defines a concept of [object locking](https://docs.aws.amazon.com/AmazonS3/latest/userguide/object-lock.html) that can be achieved either through a Retention period or a Legal hold.
| Endpoint | Garage | [Openstack Swift](https://docs.openstack.org/swift/latest/s3_compat.html) | [Ceph Object Gateway](https://docs.ceph.com/en/latest/radosgw/s3/) | [Riak CS](https://docs.riak.com/riak/cs/2.1.1/references/apis/storage/s3/index.html) | [OpenIO](https://docs.openio.io/latest/source/arch-design/s3_compliancy.html) |
| Endpoint | Garage | [Openstack Swift](https://docs.openstack.org/swift/latest/s3_compat.html) | [Ceph Object Gateway](https://docs.ceph.com/en/latest/radosgw/s3/) | [Riak CS](https://docs.riak.com/riak/cs/2.1.1/references/apis/storage/s3/index.html) | [OpenIO](https://docs.openio.io/latest/source/arch-design/s3_compliancy.html) |
|------------------------------|----------------------------------|-----------------|---------------|---------|-----|
| [GetObjectLegalHold](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetObjectLegalHold.html) | ❌ Missing | ❌| ✅ | ❌| ❌|
| [PutObjectLegalHold](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutObjectLegalHold.html) | ❌ Missing | ❌| ✅ | ❌| ❌|
| [GetObjectRetention](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetObjectRetention.html) | ❌ Missing | ❌| ✅ | ❌| ❌|
| [PutObjectRetention](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutObjectRetention.html) | ❌ Missing | ❌| ✅ | ❌| ❌|
| [GetObjectLockConfiguration](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetObjectLockConfiguration.html) | ❌ Missing | ❌| | ❌| ❌|
| [PutObjectLockConfiguration](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutObjectLockConfiguration.html) | ❌ Missing | ❌| | ❌| ❌|
| [GetObjectLockConfiguration](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetObjectLockConfiguration.html) | ❌ Missing | ❌| | ❌| ❌|
| [PutObjectLockConfiguration](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutObjectLockConfiguration.html) | ❌ Missing | ❌| | ❌| ❌|
### (Server-side) encryption
We think that you can either encrypt your server partition or do client-side encryption, so we did not implement server-side encryption for Garage.
Please open an issue if you have a use case.
| Endpoint | Garage | [Openstack Swift](https://docs.openstack.org/swift/latest/s3_compat.html) | [Ceph Object Gateway](https://docs.ceph.com/en/latest/radosgw/s3/) | [Riak CS](https://docs.riak.com/riak/cs/2.1.1/references/apis/storage/s3/index.html) | [OpenIO](https://docs.openio.io/latest/source/arch-design/s3_compliancy.html) |
| Endpoint | Garage | [Openstack Swift](https://docs.openstack.org/swift/latest/s3_compat.html) | [Ceph Object Gateway](https://docs.ceph.com/en/latest/radosgw/s3/) | [Riak CS](https://docs.riak.com/riak/cs/2.1.1/references/apis/storage/s3/index.html) | [OpenIO](https://docs.openio.io/latest/source/arch-design/s3_compliancy.html) |
|------------------------------|----------------------------------|-----------------|---------------|---------|-----|
| [DeleteBucketEncryption](https://docs.aws.amazon.com/AmazonS3/latest/API/API_DeleteBucketEncryption.html) | ❌ Missing | ❌| | ❌| ❌|
| [GetBucketEncryption](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetBucketEncryption.html) | ❌ Missing | ❌| | ❌| ❌|
| [PutBucketEncryption](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutBucketEncryption.html) | ❌ Missing | ❌| | ❌| ❌|
| [DeleteBucketEncryption](https://docs.aws.amazon.com/AmazonS3/latest/API/API_DeleteBucketEncryption.html) | ❌ Missing | ❌| | ❌| ❌|
| [GetBucketEncryption](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetBucketEncryption.html) | ❌ Missing | ❌| | ❌| ❌|
| [PutBucketEncryption](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutBucketEncryption.html) | ❌ Missing | ❌| | ❌| ❌|
### Misc endpoints
| Endpoint | Garage | [Openstack Swift](https://docs.openstack.org/swift/latest/s3_compat.html) | [Ceph Object Gateway](https://docs.ceph.com/en/latest/radosgw/s3/) | [Riak CS](https://docs.riak.com/riak/cs/2.1.1/references/apis/storage/s3/index.html) | [OpenIO](https://docs.openio.io/latest/source/arch-design/s3_compliancy.html) |
| Endpoint | Garage | [Openstack Swift](https://docs.openstack.org/swift/latest/s3_compat.html) | [Ceph Object Gateway](https://docs.ceph.com/en/latest/radosgw/s3/) | [Riak CS](https://docs.riak.com/riak/cs/2.1.1/references/apis/storage/s3/index.html) | [OpenIO](https://docs.openio.io/latest/source/arch-design/s3_compliancy.html) |
|------------------------------|----------------------------------|-----------------|---------------|---------|-----|
| [GetBucketNotificationConfiguration](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetBucketNotificationConfiguration.html) | ❌ Missing | ❌| ✅ | ❌| ❌|
| [PutBucketNotificationConfiguration](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutBucketNotificationConfiguration.html) | ❌ Missing | ❌| ✅ | ❌| ❌|
| [DeleteBucketTagging](https://docs.aws.amazon.com/AmazonS3/latest/API/API_DeleteBucketTagging.html) | ❌ Missing | ❌| | ❌| ✅ |
| [GetBucketTagging](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetBucketTagging.html) | ❌ Missing | ❌| | ❌| ✅ |
| [PutBucketTagging](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutBucketTagging.html) | ❌ Missing | ❌| | ❌| ✅ |
| [DeleteObjectTagging](https://docs.aws.amazon.com/AmazonS3/latest/API/API_DeleteObjectTagging.html) | ❌ Missing | ❌| | ❌| ✅ |
| [GetObjectTagging](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetObjectTagging.html) | ❌ Missing | ❌| | ❌| ✅ |
| [PutObjectTagging](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutObjectTagging.html) | ❌ Missing | ❌| | ❌| ✅ |
| [GetObjectTorrent](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetObjectTorrent.html) | ❌ Missing | ❌| | ❌| ❌|
| [DeleteBucketTagging](https://docs.aws.amazon.com/AmazonS3/latest/API/API_DeleteBucketTagging.html) | ❌ Missing | ❌| | ❌| ✅ |
| [GetBucketTagging](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetBucketTagging.html) | ❌ Missing | ❌| | ❌| ✅ |
| [PutBucketTagging](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutBucketTagging.html) | ❌ Missing | ❌| | ❌| ✅ |
| [DeleteObjectTagging](https://docs.aws.amazon.com/AmazonS3/latest/API/API_DeleteObjectTagging.html) | ❌ Missing | ❌| | ❌| ✅ |
| [GetObjectTagging](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetObjectTagging.html) | ❌ Missing | ❌| | ❌| ✅ |
| [PutObjectTagging](https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutObjectTagging.html) | ❌ Missing | ❌| | ❌| ✅ |
| [GetObjectTorrent](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetObjectTorrent.html) | ❌ Missing | ❌| | ❌| ❌|
### Vendor specific endpoints
<details><summary>Display Amazon specifc endpoints</summary>
| Endpoint | Garage | [Openstack Swift](https://docs.openstack.org/swift/latest/s3_compat.html) | [Ceph Object Gateway](https://docs.ceph.com/en/latest/radosgw/s3/) | [Riak CS](https://docs.riak.com/riak/cs/2.1.1/references/apis/storage/s3/index.html) | [OpenIO](https://docs.openio.io/latest/source/arch-design/s3_compliancy.html) |
| Endpoint | Garage | [Openstack Swift](https://docs.openstack.org/swift/latest/s3_compat.html) | [Ceph Object Gateway](https://docs.ceph.com/en/latest/radosgw/s3/) | [Riak CS](https://docs.riak.com/riak/cs/2.1.1/references/apis/storage/s3/index.html) | [OpenIO](https://docs.openio.io/latest/source/arch-design/s3_compliancy.html) |
|------------------------------|----------------------------------|-----------------|---------------|---------|-----|
| [DeleteBucketAnalyticsConfiguration](https://docs.aws.amazon.com/AmazonS3/latest/API/API_DeleteBucketAnalyticsConfiguration.html) | ❌ Missing | ❌| ❌| ❌| ❌|
| [DeleteBucketIntelligentTieringConfiguration](https://docs.aws.amazon.com/AmazonS3/latest/API/API_DeleteBucketIntelligentTieringConfiguration.html) | ❌ Missing | ❌| ❌| ❌| ❌|
+14
View File
@@ -0,0 +1,14 @@
*
!assets
!.gitignore
!*.svg
!*.png
!*.jpg
!*.tex
!Makefile
!.gitignore
!assets/*.drawio.pdf
!talk.pdf
+5
View File
@@ -0,0 +1,5 @@
talk.pdf: talk.tex assets/consistent_hashing_1.pdf assets/consistent_hashing_2.pdf assets/consistent_hashing_3.pdf assets/consistent_hashing_4.pdf assets/garage_tables.pdf assets/deuxfleurs.pdf
pdflatex talk.tex
assets/%.pdf: assets/%.svg
inkscape -D -z --file=$^ --export-pdf=$@
Binary file not shown.

After

Width:  |  Height:  |  Size: 32 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 115 KiB

File diff suppressed because one or more lines are too long

After

Width:  |  Height:  |  Size: 184 KiB

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:9198d0cfc0e04a56f84a353dd660b294ecf0dc68fe429892c8e9e604de016748
size 31966
Binary file not shown.

After

Width:  |  Height:  |  Size: 26 KiB

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:14dbf4a45545889babb2fdc6580399811366f6cc786cb28e3467cbcedbfe9482
size 31688
Binary file not shown.

After

Width:  |  Height:  |  Size: 27 KiB

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:21a3bb004ad35c8b47f14195c720048f8db31f47bd6346c1d747000570149c67
size 31073
Binary file not shown.

After

Width:  |  Height:  |  Size: 8.9 KiB

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:c4e109078a784c3999ca2ccd1adda8ff5793d1b230a9e27b4ad2e419afa0d37d
size 25145
Binary file not shown.

After

Width:  |  Height:  |  Size: 18 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 4.8 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 263 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 82 KiB

File diff suppressed because one or more lines are too long

After

Width:  |  Height:  |  Size: 53 KiB

File diff suppressed because one or more lines are too long

After

Width:  |  Height:  |  Size: 54 KiB

File diff suppressed because one or more lines are too long

After

Width:  |  Height:  |  Size: 56 KiB

File diff suppressed because one or more lines are too long

After

Width:  |  Height:  |  Size: 57 KiB

@@ -0,0 +1,91 @@
<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<svg
viewBox="0 0 70.424515 70.300102"
version="1.1"
id="svg8"
sodipodi:docname="logo.svg"
inkscape:version="1.1 (c68e22c387, 2021-05-23)"
inkscape:export-filename="/home/quentin/Documents/dev/deuxfleurs/site/src/img/logo.png"
inkscape:export-xdpi="699.30194"
inkscape:export-ydpi="699.30194"
width="70.424515"
height="70.300102"
xmlns:inkscape="http://www.inkscape.org/namespaces/inkscape"
xmlns:sodipodi="http://sodipodi.sourceforge.net/DTD/sodipodi-0.dtd"
xmlns="http://www.w3.org/2000/svg"
xmlns:svg="http://www.w3.org/2000/svg">
<defs
id="defs12" />
<sodipodi:namedview
id="namedview10"
pagecolor="#ffffff"
bordercolor="#666666"
borderopacity="1.0"
inkscape:pageshadow="2"
inkscape:pageopacity="0.0"
inkscape:pagecheckerboard="0"
showgrid="false"
inkscape:zoom="12.125"
inkscape:cx="43.092783"
inkscape:cy="48.082474"
inkscape:window-width="3072"
inkscape:window-height="1659"
inkscape:window-x="0"
inkscape:window-y="0"
inkscape:window-maximized="1"
inkscape:current-layer="svg8" />
<g
id="g79969"
transform="translate(-0.827,34.992103)">
<path
fill="#ffffff"
d="m 15.632,34.661 c -0.799,-0.597 -1.498,-1.484 -2.035,-2.592 l -0.228,-0.47 -0.46,0.249 c -0.975,0.528 -1.913,0.858 -2.744,0.969 L 9.963,29.061 6.327,30.029 C 6.17,29.175 6.202,28.142 6.423,27.007 L 6.526,26.482 5.994,26.416 C 4.752,26.262 3.688,25.891 2.89,25.336 L 4.411,22.419 1.423,20.896 C 1.742,19.952 2.371,19.014 3.257,18.161 L 3.634,17.798 3.255,17.438 C 2.452,16.674 1.847,15.884 1.485,15.127 L 4.995,13.774 2.95,10.615 C 3.69,10.213 4.643,9.929 5.739,9.783 L 6.258,9.715 6.167,9.201 C 5.952,7.99 5.995,6.863 6.291,5.913 l 3.308,0.523 0.524,-3.308 c 0.988,0.013 2.08,0.326 3.164,0.907 L 13.749,4.283 13.975,3.81 C 14.454,2.807 15.019,1.986 15.628,1.406 L 18,4.326 20.372,1.406 c 0.609,0.58 1.175,1.401 1.653,2.404 l 0.226,0.473 0.462,-0.247 C 23.798,3.455 24.891,3.142 25.877,3.13 L 26.4,6.438 29.71,5.913 c 0.296,0.951 0.34,2.078 0.124,3.288 l -0.092,0.515 0.518,0.069 c 1.095,0.145 2.048,0.43 2.788,0.832 l -2.046,3.156 3.511,1.355 c -0.361,0.757 -0.966,1.547 -1.77,2.311 l -0.379,0.36 0.377,0.363 c 0.888,0.854 1.516,1.793 1.835,2.736 l -2.984,1.52 1.521,2.984 c -0.812,0.574 -1.871,0.964 -3.094,1.134 l -0.518,0.072 0.096,0.514 c 0.201,1.089 0.226,2.083 0.073,2.909 l -3.634,-0.97 -0.204,3.757 c -0.83,-0.11 -1.768,-0.44 -2.742,-0.968 l -0.459,-0.249 -0.228,0.47 c -0.539,1.107 -1.237,1.994 -2.036,2.591 L 18,32.293 Z"
id="path2" />
<path
d="M 7.092,10.678 C 6.562,9.189 6.394,7.708 6.66,6.478 l 2.368,0.375 0.987,0.156 0.157,-0.988 0.375,-2.368 C 11.808,3.78 13.16,4.396 14.409,5.359 14.527,5.022 14.653,4.696 14.791,4.392 13.24,3.257 11.568,2.629 10.061,2.629 9.938,2.629 9.816,2.633 9.695,2.642 L 9.184,5.865 5.96,5.354 C 5.36,6.841 5.395,8.769 6.045,10.747 6.38,10.71 6.729,10.686 7.092,10.678 Z M 21.593,5.359 c 1.248,-0.962 2.6,-1.578 3.86,-1.705 l 0.376,2.368 0.156,0.988 0.987,-0.157 2.369,-0.376 c 0.266,1.23 0.098,2.71 -0.432,4.2 0.361,0.009 0.711,0.032 1.046,0.07 C 30.606,8.769 30.64,6.841 30.04,5.353 L 26.815,5.865 26.304,2.641 c -0.12,-0.008 -0.242,-0.012 -0.365,-0.012 -1.507,0 -3.179,0.628 -4.73,1.762 0.14,0.306 0.266,0.631 0.384,0.968 z M 7.368,27 h 0.035 c 0.067,0 0.157,-0.604 0.26,-0.947 -0.098,0.004 -0.197,0.046 -0.294,0.046 -1.496,0 -2.826,-0.303 -3.83,-0.89 L 4.628,23.081 5.082,22.194 4.191,21.742 2.055,20.654 C 2.563,19.503 3.57,18.404 4.873,17.511 4.586,17.292 4.312,17.07 4.063,16.842 2.376,18.059 1.217,19.597 0.828,21.152 l 2.908,1.483 -1.482,2.843 C 3.475,26.501 5.303,27 7.368,27 Z m 27.806,-5.846 c -0.39,-1.555 -1.548,-3.093 -3.234,-4.311 -0.25,0.228 -0.523,0.451 -0.81,0.669 1.304,0.893 2.31,1.992 2.817,3.145 l -2.136,1.088 -0.891,0.453 0.454,0.892 1.089,2.137 c -1.004,0.587 -2.332,0.904 -3.828,0.904 -0.099,0 -0.199,-0.01 -0.299,-0.013 0.103,0.344 0.192,0.683 0.26,1.011 l 0.039,0.002 c 2.066,0 3.892,-0.563 5.112,-1.587 l -1.482,-2.908 z m -12.653,9.182 c -0.447,1.517 -1.181,2.812 -2.119,3.651 L 18.707,32.293 18,31.586 l -0.707,0.707 -1.695,1.694 c -0.938,-0.839 -1.673,-2.136 -2.12,-3.652 -0.296,0.206 -0.593,0.397 -0.886,0.563 0.636,1.98 1.741,3.559 3.1,4.409 L 18,33 l 2.308,2.308 c 1.358,-0.851 2.464,-2.428 3.101,-4.408 -0.295,-0.168 -0.591,-0.359 -0.888,-0.564 z"
fill="#ea596e"
id="path4" />
<path
fill="#ea596e"
d="m 20.118,5.683 c 0.426,1.146 0.748,2.596 0.841,4.284 l 0.2,3.683 3.564,-0.946 c 1.32,-0.351 2.655,-0.536 3.86,-0.536 0.16,0 0.318,0.003 0.474,0.01 l -1.827,2.819 3.139,1.211 c -0.958,0.759 -2.237,1.514 -3.814,2.123 l -3.441,1.328 2.001,3.099 c 0.918,1.42 1.509,2.782 1.838,3.96 L 23.709,25.853 23.527,29.21 C 22.508,28.533 21.395,27.55 20.329,26.237 L 18,23.374 15.672,26.236 c -1.066,1.312 -2.179,2.295 -3.198,2.972 l -0.18,-3.354 -3.248,0.864 c 0.329,-1.178 0.921,-2.54 1.839,-3.961 L 12.889,19.658 9.447,18.33 C 7.87,17.721 6.591,16.967 5.633,16.208 L 8.768,15 6.941,12.177 c 0.155,-0.006 0.313,-0.01 0.473,-0.01 1.206,0 2.541,0.185 3.861,0.536 l 3.564,0.947 0.202,-3.683 c 0.092,-1.688 0.415,-3.138 0.84,-4.284 L 18,8.292 20.118,5.683 M 20.308,0.692 18,3.533 15.692,0.692 C 13.703,2.224 12.271,5.684 12.046,9.804 10.429,9.374 8.854,9.167 7.414,9.167 c -2.11,0 -3.929,0.445 -5.161,1.289 l 1.989,3.073 -3.415,1.316 c 0.842,2.366 3.69,4.797 7.54,6.283 -2.241,3.465 -3.116,7.106 -2.407,9.516 l 3.537,-0.941 0.196,3.654 c 2.512,-0.07 5.703,-2.027 8.307,-5.228 2.603,3.201 5.796,5.158 8.306,5.228 l 0.198,-3.655 3.535,0.943 c 0.71,-2.411 -0.165,-6.05 -2.404,-9.517 3.849,-1.485 6.696,-3.918 7.538,-6.283 l -3.415,-1.318 1.99,-3.07 c -1.233,-0.844 -3.053,-1.29 -5.164,-1.29 -1.438,0 -3.013,0.207 -4.63,0.636 C 23.729,5.684 22.297,2.224 20.308,0.692 Z"
id="path6" />
</g>
<g
id="g79964"
transform="translate(-1.043816,35.993714)">
<path
fill="#ffffff"
d="m 51.92633,-2.0247139 c -0.799,-0.597 -1.498,-1.484 -2.035,-2.592 l -0.228,-0.47 -0.46,0.249 c -0.975,0.528 -1.913,0.858 -2.744,0.969 l -0.202,-3.7560001 -3.636,0.968 c -0.157,-0.854 -0.125,-1.887 0.096,-3.022 l 0.103,-0.525 -0.532,-0.066 c -1.242,-0.154 -2.306,-0.525 -3.104,-1.08 l 1.521,-2.917 -2.988,-1.523 c 0.319,-0.944 0.948,-1.882 1.834,-2.735 l 0.377,-0.363 -0.379,-0.36 c -0.803,-0.764 -1.408,-1.554 -1.77,-2.311 l 3.51,-1.353 -2.045,-3.159 c 0.74,-0.402 1.693,-0.686 2.789,-0.832 l 0.519,-0.068 -0.091,-0.514 c -0.215,-1.211 -0.172,-2.338 0.124,-3.288 l 3.308,0.523 0.524,-3.308 c 0.988,0.013 2.08,0.326 3.164,0.907 l 0.462,0.248 0.226,-0.473 c 0.479,-1.003 1.044,-1.824 1.653,-2.404 l 2.372,2.92 2.372,-2.92 c 0.609,0.58 1.175,1.401 1.653,2.404 l 0.226,0.473 0.462,-0.247 c 1.085,-0.581 2.178,-0.894 3.164,-0.906 l 0.523,3.308 3.31,-0.525 c 0.296,0.951 0.34,2.078 0.124,3.288 l -0.092,0.515 0.518,0.069 c 1.095,0.145 2.048,0.43 2.788,0.832 l -2.046,3.156 3.511,1.355 c -0.361,0.757 -0.966,1.547 -1.77,2.311 l -0.379,0.36 0.377,0.363 c 0.888,0.854 1.516,1.793 1.835,2.736 l -2.984,1.52 1.521,2.984 c -0.812,0.574 -1.871,0.964 -3.094,1.134 l -0.518,0.072 0.096,0.514 c 0.201,1.089 0.226,2.083 0.073,2.909 l -3.634,-0.97 -0.204,3.7570001 c -0.83,-0.11 -1.768,-0.44 -2.742,-0.968 l -0.459,-0.249 -0.228,0.47 c -0.539,1.107 -1.237,1.994 -2.036,2.591 l -2.367,-2.369 z"
id="path2-9" />
<path
d="m 43.38633,-26.007714 c -0.53,-1.489 -0.698,-2.97 -0.432,-4.2 l 2.368,0.375 0.987,0.156 0.157,-0.988 0.375,-2.368 c 1.261,0.127 2.613,0.743 3.862,1.706 0.118,-0.337 0.244,-0.663 0.382,-0.967 -1.551,-1.135 -3.223,-1.763 -4.73,-1.763 -0.123,0 -0.245,0.004 -0.366,0.013 l -0.511,3.223 -3.224,-0.511 c -0.6,1.487 -0.565,3.415 0.085,5.393 0.335,-0.037 0.684,-0.061 1.047,-0.069 z m 14.501,-5.319 c 1.248,-0.962 2.6,-1.578 3.86,-1.705 l 0.376,2.368 0.156,0.988 0.987,-0.157 2.369,-0.376 c 0.266,1.23 0.098,2.71 -0.432,4.2 0.361,0.009 0.711,0.032 1.046,0.07 0.651,-1.978 0.685,-3.906 0.085,-5.394 l -3.225,0.512 -0.511,-3.224 c -0.12,-0.008 -0.242,-0.012 -0.365,-0.012 -1.507,0 -3.179,0.628 -4.73,1.762 0.14,0.306 0.266,0.631 0.384,0.968 z m -14.225,21.641 h 0.035 c 0.067,0 0.157,-0.604 0.26,-0.947 -0.098,0.004 -0.197,0.046 -0.294,0.046 -1.496,0 -2.826,-0.303 -3.83,-0.89 l 1.089,-2.128 0.454,-0.887 -0.891,-0.452 -2.136,-1.088 c 0.508,-1.151 1.515,-2.25 2.818,-3.143 -0.287,-0.219 -0.561,-0.441 -0.81,-0.669 -1.687,1.217 -2.846,2.755 -3.235,4.31 l 2.908,1.483 -1.482,2.843 c 1.221,1.023 3.049,1.522 5.114,1.522 z m 27.806,-5.846 c -0.39,-1.555 -1.548,-3.093 -3.234,-4.311 -0.25,0.228 -0.523,0.451 -0.81,0.669 1.304,0.893 2.31,1.992 2.817,3.145 l -2.136,1.088 -0.891,0.453 0.454,0.892 1.089,2.137 c -1.004,0.587 -2.332,0.904 -3.828,0.904 -0.099,0 -0.199,-0.01 -0.299,-0.013 0.103,0.344 0.192,0.683 0.26,1.011 l 0.039,0.002 c 2.066,0 3.892,-0.563 5.112,-1.587 l -1.482,-2.908 z m -12.653,9.182 c -0.447,1.5170001 -1.181,2.8120001 -2.119,3.6510001 l -1.695,-1.694 -0.707,-0.707 -0.707,0.707 -1.695,1.694 c -0.938,-0.839 -1.673,-2.136 -2.12,-3.6520001 -0.296,0.2060001 -0.593,0.3970001 -0.886,0.5630001 0.636,1.98 1.741,3.559 3.1,4.409 l 2.308,-2.307 2.308,2.308 c 1.358,-0.851 2.464,-2.428 3.101,-4.408 -0.295,-0.168 -0.591,-0.359 -0.888,-0.5640001 z"
fill="#ea596e"
id="path4-3" />
<path
fill="#ea596e"
d="m 56.41233,-31.002714 c 0.426,1.146 0.748,2.596 0.841,4.284 l 0.2,3.683 3.564,-0.946 c 1.32,-0.351 2.655,-0.536 3.86,-0.536 0.16,0 0.318,0.003 0.474,0.01 l -1.827,2.819 3.139,1.211 c -0.958,0.759 -2.237,1.514 -3.814,2.123 l -3.441,1.328 2.001,3.099 c 0.918,1.42 1.509,2.782 1.838,3.96 l -3.244,-0.865 -0.182,3.357 c -1.019,-0.677 -2.132,-1.66 -3.198,-2.973 l -2.329,-2.863 -2.328,2.862 c -1.066,1.312 -2.179,2.295 -3.198,2.972 l -0.18,-3.354 -3.248,0.864 c 0.329,-1.178 0.921,-2.54 1.839,-3.961 l 2.004,-3.099 -3.442,-1.328 c -1.577,-0.609 -2.856,-1.363 -3.814,-2.122 l 3.135,-1.208 -1.827,-2.823 c 0.155,-0.006 0.313,-0.01 0.473,-0.01 1.206,0 2.541,0.185 3.861,0.536 l 3.564,0.947 0.202,-3.683 c 0.092,-1.688 0.415,-3.138 0.84,-4.284 l 2.119,2.609 2.118,-2.609 m 0.19,-4.991 -2.308,2.841 -2.308,-2.841 c -1.989,1.532 -3.421,4.992 -3.646,9.112 -1.617,-0.43 -3.192,-0.637 -4.632,-0.637 -2.11,0 -3.929,0.445 -5.161,1.289 l 1.989,3.073 -3.415,1.316 c 0.842,2.366 3.69,4.797 7.54,6.283 -2.241,3.465 -3.116,7.106 -2.407,9.5160001 l 3.537,-0.9410001 0.196,3.6540001 c 2.512,-0.07 5.703,-2.027 8.307,-5.2280001 2.603,3.2010001 5.796,5.1580001 8.306,5.2280001 l 0.198,-3.6550001 3.535,0.9430001 c 0.71,-2.4110001 -0.165,-6.0500001 -2.404,-9.5170001 3.849,-1.485 6.696,-3.918 7.538,-6.283 l -3.415,-1.318 1.99,-3.07 c -1.233,-0.844 -3.053,-1.29 -5.164,-1.29 -1.438,0 -3.013,0.207 -4.63,0.636 -0.225,-4.119 -1.657,-7.579 -3.646,-9.111 z"
id="path6-6" />
</g>
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:42.6667px;line-height:1.25;font-family:sans-serif;fill:#ea596e;fill-opacity:1;stroke:none"
x="2.2188232"
y="31.430677"
id="text46212"><tspan
sodipodi:role="line"
id="tspan46210"
x="2.2188232"
y="31.430677"
style="font-style:normal;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:42.6667px;font-family:'TeX Gyre Termes';-inkscape-font-specification:'TeX Gyre Termes'">D</tspan></text>
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:42.6667px;line-height:1.25;font-family:sans-serif;fill:#ea596e;fill-opacity:1;stroke:none"
x="41.347008"
y="67.114784"
id="text46212-1"><tspan
sodipodi:role="line"
id="tspan46210-5"
x="41.347008"
y="67.114784"
style="font-style:normal;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:42.6667px;font-family:'TeX Gyre Termes';-inkscape-font-specification:'TeX Gyre Termes'">F</tspan></text>
</svg>

After

Width:  |  Height:  |  Size: 12 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 129 KiB

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:0d9ca1a04f943664e14d0e1edbde596789e4798f8ad5e1cff68bc9dd0cc1334f
size 26098
Binary file not shown.

After

Width:  |  Height:  |  Size: 13 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 88 KiB

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:bffa5d4a72ad25e0b18b43022258a2e4471ce94ff5397aab892fe43ec9d4d7d5
size 33911
@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:9c4913b52e84b5af3f8effbcb7e2060845a8c6f43d5b38072b9ee33c0300d49f
size 31051
@@ -0,0 +1,537 @@
<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<svg
width="850"
height="480"
viewBox="0 0 224.89584 127"
version="1.1"
id="svg8"
inkscape:version="1.2 (dc2aedaf03, 2022-05-15)"
sodipodi:docname="garage_tables.svg"
xmlns:inkscape="http://www.inkscape.org/namespaces/inkscape"
xmlns:sodipodi="http://sodipodi.sourceforge.net/DTD/sodipodi-0.dtd"
xmlns="http://www.w3.org/2000/svg"
xmlns:svg="http://www.w3.org/2000/svg"
xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
xmlns:cc="http://creativecommons.org/ns#"
xmlns:dc="http://purl.org/dc/elements/1.1/">
<defs
id="defs2">
<marker
style="overflow:visible"
id="marker1262"
refX="0"
refY="0"
orient="auto"
inkscape:stockid="Arrow1Mend"
inkscape:isstock="true">
<path
transform="matrix(-0.4,0,0,-0.4,-4,0)"
style="fill:#000000;fill-opacity:1;fill-rule:evenodd;stroke:#000000;stroke-width:1pt;stroke-opacity:1"
d="M 0,0 5,-5 -12.5,0 5,5 Z"
id="path1260" />
</marker>
<marker
style="overflow:visible"
id="Arrow1Mend"
refX="0"
refY="0"
orient="auto"
inkscape:stockid="Arrow1Mend"
inkscape:isstock="true"
inkscape:collect="always">
<path
transform="matrix(-0.4,0,0,-0.4,-4,0)"
style="fill:#000000;fill-opacity:1;fill-rule:evenodd;stroke:#000000;stroke-width:1pt;stroke-opacity:1"
d="M 0,0 5,-5 -12.5,0 5,5 Z"
id="path965" />
</marker>
<marker
style="overflow:visible"
id="Arrow1Lend"
refX="0"
refY="0"
orient="auto"
inkscape:stockid="Arrow1Lend"
inkscape:isstock="true">
<path
transform="matrix(-0.8,0,0,-0.8,-10,0)"
style="fill:#000000;fill-opacity:1;fill-rule:evenodd;stroke:#000000;stroke-width:1pt;stroke-opacity:1"
d="M 0,0 5,-5 -12.5,0 5,5 Z"
id="path959" />
</marker>
</defs>
<sodipodi:namedview
id="base"
pagecolor="#ffffff"
bordercolor="#666666"
borderopacity="1.0"
inkscape:pageopacity="0.0"
inkscape:pageshadow="2"
inkscape:zoom="0.98994949"
inkscape:cx="429.31483"
inkscape:cy="289.40871"
inkscape:document-units="mm"
inkscape:current-layer="layer1"
inkscape:document-rotation="0"
showgrid="false"
units="px"
inkscape:window-width="1678"
inkscape:window-height="993"
inkscape:window-x="0"
inkscape:window-y="0"
inkscape:window-maximized="1"
inkscape:showpageshadow="2"
inkscape:pagecheckerboard="0"
inkscape:deskcolor="#d1d1d1" />
<metadata
id="metadata5">
<rdf:RDF>
<cc:Work
rdf:about="">
<dc:format>image/svg+xml</dc:format>
<dc:type
rdf:resource="http://purl.org/dc/dcmitype/StillImage" />
<dc:title />
</cc:Work>
</rdf:RDF>
</metadata>
<g
inkscape:label="Layer 1"
inkscape:groupmode="layer"
id="layer1">
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="39.570904"
y="38.452755"
id="text2025"><tspan
sodipodi:role="line"
id="tspan2023"
x="39.570904"
y="38.452755"
style="font-size:5.64444px;stroke-width:0.264583" /></text>
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:10.5833px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="101.95796"
y="92.835831"
id="text2139"><tspan
sodipodi:role="line"
id="tspan2137"
x="101.95796"
y="92.835831"
style="stroke-width:0.264583"> </tspan></text>
<g
id="g2316"
transform="translate(-11.455511,1.5722486)">
<g
id="g2277">
<rect
style="fill:none;stroke:#000000;stroke-width:0.8;stroke-miterlimit:4;stroke-dasharray:none;stroke-dashoffset:0;stroke-opacity:1"
id="rect833"
width="47.419891"
height="95.353409"
x="18.534418"
y="24.42766" />
<rect
style="fill:none;stroke:#000000;stroke-width:0.799999;stroke-miterlimit:4;stroke-dasharray:none;stroke-dashoffset:0;stroke-opacity:1"
id="rect833-3"
width="47.419891"
height="86.973076"
x="18.534418"
y="32.807987" />
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="32.250839"
y="29.894743"
id="text852"><tspan
sodipodi:role="line"
id="tspan850"
x="32.250839"
y="29.894743"
style="font-style:normal;font-variant:normal;font-weight:bold;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';stroke-width:0.264583">Object</tspan></text>
</g>
<g
id="g2066"
transform="translate(-2.1807817,-3.0621439)">
<g
id="g1969"
transform="matrix(0.12763631,0,0,0.12763631,0.7215051,24.717273)"
style="fill:#ff6600;fill-opacity:1;stroke:none;stroke-opacity:1">
<path
style="fill:#ff6600;fill-opacity:1;stroke:none;stroke-width:0.264583;stroke-opacity:1"
d="m 203.71837,154.80038 c -1.11451,3.75057 -2.45288,5.84095 -5.11132,7.98327 -2.2735,1.83211 -4.66721,2.65982 -8.09339,2.79857 -2.59227,0.10498 -2.92868,0.0577 -5.02863,-0.70611 -3.99215,-1.45212 -7.1627,-4.65496 -8.48408,-8.57046 -1.28374,-3.80398 -0.61478,-8.68216 1.64793,-12.01698 0.87317,-1.28689 3.15089,-3.48326 4.18771,-4.03815 l 0.53332,-28.51234 5.78454,-5.09197 6.95158,6.16704 -3.21112,3.49026 3.17616,3.45499 -3.17616,3.40822 2.98973,3.28645 -3.24843,3.3829 4.49203,4.58395 0.0516,5.69106 c 1.06874,0.64848 3.81974,3.24046 4.69548,4.56257 0.452,0.68241 1.06834,2.0197 1.36962,2.97176 0.62932,1.98864 0.88051,5.785 0.47342,7.15497 z m -10.0406,2.32604 c -0.88184,-3.17515 -4.92402,-3.78864 -6.75297,-1.02492 -0.58328,0.8814 -0.6898,1.28852 -0.58362,2.23056 0.26492,2.35041 2.45434,3.95262 4.60856,3.37255 1.19644,-0.32217 2.39435,-1.44872 2.72875,-2.56621 0.30682,-1.02529 0.30686,-0.9045 -7.9e-4,-2.01198 z"
id="path1971"
sodipodi:nodetypes="ssscsscccccccccccssscsssscc" />
</g>
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="28.809687"
y="44.070885"
id="text852-9"><tspan
sodipodi:role="line"
id="tspan850-4"
x="28.809687"
y="44.070885"
style="font-style:normal;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';stroke-width:0.264583">bucket </tspan></text>
</g>
<g
id="g2066-7"
transform="translate(-2.1807817,6.2627616)">
<g
id="g1969-8"
transform="matrix(0.12763631,0,0,0.12763631,0.7215051,24.717273)"
style="fill:#ff6600;fill-opacity:1;stroke:none;stroke-opacity:1">
<path
style="fill:#4040ff;fill-opacity:1;stroke:none;stroke-width:0.264583;stroke-opacity:1"
d="m 203.71837,154.80038 c -1.11451,3.75057 -2.45288,5.84095 -5.11132,7.98327 -2.2735,1.83211 -4.66721,2.65982 -8.09339,2.79857 -2.59227,0.10498 -2.92868,0.0577 -5.02863,-0.70611 -3.99215,-1.45212 -7.1627,-4.65496 -8.48408,-8.57046 -1.28374,-3.80398 -0.61478,-8.68216 1.64793,-12.01698 0.87317,-1.28689 3.15089,-3.48326 4.18771,-4.03815 l 0.53332,-28.51234 5.78454,-5.09197 6.95158,6.16704 -3.21112,3.49026 3.17616,3.45499 -3.17616,3.40822 2.98973,3.28645 -3.24843,3.3829 4.49203,4.58395 0.0516,5.69106 c 1.06874,0.64848 3.81974,3.24046 4.69548,4.56257 0.452,0.68241 1.06834,2.0197 1.36962,2.97176 0.62932,1.98864 0.88051,5.785 0.47342,7.15497 z m -10.0406,2.32604 c -0.88184,-3.17515 -4.92402,-3.78864 -6.75297,-1.02492 -0.58328,0.8814 -0.6898,1.28852 -0.58362,2.23056 0.26492,2.35041 2.45434,3.95262 4.60856,3.37255 1.19644,-0.32217 2.39435,-1.44872 2.72875,-2.56621 0.30682,-1.02529 0.30686,-0.9045 -7.9e-4,-2.01198 z"
id="path1971-4"
sodipodi:nodetypes="ssscsscccccccccccssscsssscc" />
</g>
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="28.809687"
y="44.070885"
id="text852-9-5"><tspan
sodipodi:role="line"
id="tspan850-4-0"
x="28.809687"
y="44.070885"
style="font-style:normal;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';stroke-width:0.264583">file path </tspan></text>
<path
style="fill:#ff6600;fill-opacity:1;stroke:none;stroke-width:0.0337704;stroke-opacity:1"
d="m 174.20027,104.45585 c -0.14225,0.47871 -0.31308,0.74552 -0.65239,1.01896 -0.29018,0.23384 -0.5957,0.33949 -1.03301,0.3572 -0.33087,0.0134 -0.37381,0.007 -0.64184,-0.0901 -0.50954,-0.18534 -0.91422,-0.59414 -1.08287,-1.0939 -0.16385,-0.48552 -0.0785,-1.10816 0.21033,-1.5338 0.11145,-0.16426 0.40217,-0.44459 0.53451,-0.51542 l 0.0681,-3.639207 0.73832,-0.64992 0.88727,0.787138 -0.40986,0.445484 0.4054,0.440982 -0.4054,0.435013 0.3816,0.41947 -0.41461,0.43178 0.57334,0.58508 0.007,0.72639 c 0.13641,0.0828 0.48753,0.4136 0.59931,0.58235 0.0577,0.0871 0.13636,0.25778 0.17481,0.3793 0.0803,0.25382 0.11239,0.73838 0.0604,0.91323 z m -1.28154,0.29689 c -0.11256,-0.40526 -0.62849,-0.48357 -0.86193,-0.13082 -0.0745,0.1125 -0.088,0.16447 -0.0745,0.2847 0.0338,0.3 0.31326,0.5045 0.58822,0.43046 0.15271,-0.0411 0.30561,-0.1849 0.34829,-0.32754 0.0392,-0.13086 0.0392,-0.11544 -1e-4,-0.2568 z"
id="path1971-3"
sodipodi:nodetypes="ssscsscccccccccccssscsssscc" />
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="177.8474"
y="104.05132"
id="text852-9-6"><tspan
sodipodi:role="line"
id="tspan850-4-7"
x="177.8474"
y="104.05132"
style="font-style:normal;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';stroke-width:0.264583">= partition key </tspan></text>
<path
style="fill:#4040ff;fill-opacity:1;stroke:none;stroke-width:0.0337704;stroke-opacity:1"
d="m 174.20027,113.78076 c -0.14225,0.47871 -0.31308,0.74552 -0.65239,1.01895 -0.29018,0.23385 -0.5957,0.33949 -1.03301,0.3572 -0.33087,0.0134 -0.37381,0.007 -0.64184,-0.0901 -0.50954,-0.18534 -0.91422,-0.59414 -1.08287,-1.0939 -0.16385,-0.48553 -0.0785,-1.10816 0.21033,-1.53381 0.11145,-0.16425 0.40217,-0.44459 0.53451,-0.51541 l 0.0681,-3.63921 0.73832,-0.64992 0.88727,0.78714 -0.40986,0.44548 0.4054,0.44098 -0.4054,0.43502 0.3816,0.41947 -0.41461,0.43178 0.57334,0.58508 0.007,0.72638 c 0.13641,0.0828 0.48753,0.4136 0.59931,0.58235 0.0577,0.0871 0.13636,0.25779 0.17481,0.37931 0.0803,0.25382 0.11239,0.73837 0.0604,0.91323 z m -1.28154,0.29689 c -0.11256,-0.40527 -0.62849,-0.48357 -0.86193,-0.13082 -0.0745,0.1125 -0.088,0.16446 -0.0745,0.2847 0.0338,0.3 0.31326,0.5045 0.58822,0.43046 0.15271,-0.0411 0.30561,-0.18491 0.34829,-0.32754 0.0392,-0.13087 0.0392,-0.11545 -1e-4,-0.2568 z"
id="path1971-4-5"
sodipodi:nodetypes="ssscsscccccccccccssscsssscc" />
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="177.8474"
y="113.37622"
id="text852-9-5-3"><tspan
sodipodi:role="line"
id="tspan850-4-0-5"
x="177.8474"
y="113.37622"
style="font-style:normal;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';stroke-width:0.264583">= sort key </tspan></text>
</g>
<g
id="g2161"
transform="translate(-62.264403,-59.333115)">
<g
id="g2271"
transform="translate(0,67.042823)">
<rect
style="fill:none;stroke:#000000;stroke-width:0.799999;stroke-miterlimit:4;stroke-dasharray:none;stroke-dashoffset:0;stroke-opacity:1"
id="rect833-6"
width="39.008453"
height="16.775949"
x="84.896881"
y="90.266838" />
<rect
style="fill:none;stroke:#000000;stroke-width:0.799999;stroke-miterlimit:4;stroke-dasharray:none;stroke-dashoffset:0;stroke-opacity:1"
id="rect833-3-1"
width="39.008453"
height="8.673645"
x="84.896881"
y="98.369141" />
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="89.826942"
y="96.212921"
id="text852-0"><tspan
sodipodi:role="line"
id="tspan850-6"
x="89.826942"
y="96.212921"
style="font-style:normal;font-variant:normal;font-weight:bold;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';stroke-width:0.264583">Version 1</tspan></text>
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="89.826942"
y="104.71013"
id="text852-0-3"><tspan
sodipodi:role="line"
id="tspan850-6-2"
x="89.826942"
y="104.71013"
style="font-style:italic;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';fill:#4d4d4d;stroke-width:0.264583">deleted</tspan></text>
</g>
</g>
<g
id="g2263"
transform="translate(0,-22.791204)">
<g
id="g2161-1"
transform="translate(-62.264403,-10.910843)">
<rect
style="fill:none;stroke:#000000;stroke-width:0.799999;stroke-miterlimit:4;stroke-dasharray:none;stroke-dashoffset:0;stroke-opacity:1"
id="rect833-6-5"
width="39.008453"
height="36.749603"
x="84.896881"
y="90.266838" />
<rect
style="fill:none;stroke:#000000;stroke-width:0.799999;stroke-miterlimit:4;stroke-dasharray:none;stroke-dashoffset:0;stroke-opacity:1"
id="rect833-3-1-5"
width="39.008453"
height="28.647301"
x="84.896881"
y="98.369141" />
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="89.826942"
y="96.212921"
id="text852-0-4"><tspan
sodipodi:role="line"
id="tspan850-6-7"
x="89.826942"
y="96.212921"
style="font-style:normal;font-variant:normal;font-weight:bold;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';stroke-width:0.264583">Version 2</tspan></text>
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="89.826942"
y="104.71013"
id="text852-0-3-6"><tspan
sodipodi:role="line"
id="tspan850-6-2-5"
x="89.826942"
y="104.71013"
style="font-style:normal;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';fill:#000000;stroke-width:0.264583">id</tspan></text>
</g>
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="27.56254"
y="100.34132"
id="text852-0-3-6-6"><tspan
sodipodi:role="line"
id="tspan850-6-2-5-9"
x="27.56254"
y="100.34132"
style="font-style:normal;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';fill:#000000;stroke-width:0.264583">size</tspan></text>
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="27.56254"
y="106.90263"
id="text852-0-3-6-6-3"><tspan
sodipodi:role="line"
id="tspan850-6-2-5-9-7"
x="27.56254"
y="106.90263"
style="font-style:normal;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';fill:#000000;stroke-width:0.264583">MIME type</tspan></text>
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="27.56254"
y="111.92816"
id="text852-0-3-6-6-3-4"><tspan
sodipodi:role="line"
id="tspan850-6-2-5-9-7-5"
x="27.56254"
y="111.92816"
style="font-style:normal;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';fill:#000000;stroke-width:0.264583">...</tspan></text>
</g>
</g>
<g
id="g898"
transform="translate(-6.2484318,29.95006)">
<rect
style="fill:none;stroke:#000000;stroke-width:0.799999;stroke-miterlimit:4;stroke-dasharray:none;stroke-dashoffset:0;stroke-opacity:1"
id="rect833-7"
width="47.419891"
height="44.007515"
x="95.443573"
y="24.42766" />
<rect
style="fill:none;stroke:#000000;stroke-width:0.799999;stroke-miterlimit:4;stroke-dasharray:none;stroke-dashoffset:0;stroke-opacity:1"
id="rect833-3-4"
width="47.419891"
height="35.627186"
x="95.443573"
y="32.807987" />
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="107.46638"
y="29.894743"
id="text852-4"><tspan
sodipodi:role="line"
id="tspan850-3"
x="107.46638"
y="29.894743"
style="font-style:normal;font-variant:normal;font-weight:bold;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';stroke-width:0.264583">Version</tspan></text>
<path
style="fill:#ff6600;fill-opacity:1;stroke:none;stroke-width:0.0337704;stroke-opacity:1"
d="m 102.90563,41.413279 c -0.14226,0.478709 -0.31308,0.745518 -0.65239,1.018956 -0.29019,0.233843 -0.59571,0.339489 -1.03301,0.357199 -0.33087,0.0134 -0.37381,0.0074 -0.64184,-0.09013 -0.50954,-0.185343 -0.914221,-0.594142 -1.082877,-1.093901 -0.163852,-0.485526 -0.07847,-1.108159 0.210335,-1.533803 0.111448,-0.164254 0.402172,-0.444591 0.534502,-0.515415 l 0.0681,-3.63921 0.73832,-0.64992 0.88727,0.787138 -0.40985,0.445484 0.40539,0.440982 -0.40539,0.435013 0.3816,0.41947 -0.41462,0.431781 0.57335,0.585078 0.007,0.726386 c 0.13641,0.08277 0.48753,0.413601 0.59931,0.58235 0.0577,0.0871 0.13636,0.257787 0.17481,0.379304 0.0803,0.253823 0.11239,0.738377 0.0604,0.913234 z m -1.28155,0.296888 c -0.11255,-0.405265 -0.62848,-0.483569 -0.86192,-0.130817 -0.0744,0.112498 -0.088,0.164461 -0.0745,0.2847 0.0338,0.299998 0.31326,0.504498 0.58822,0.43046 0.15271,-0.04112 0.3056,-0.184909 0.34828,-0.327542 0.0392,-0.130864 0.0392,-0.115447 -1e-4,-0.256801 z"
id="path1971-0"
sodipodi:nodetypes="ssscsscccccccccccssscsssscc" />
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="104.99195"
y="41.008743"
id="text852-9-7"><tspan
sodipodi:role="line"
id="tspan850-4-8"
x="104.99195"
y="41.008743"
style="font-style:normal;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';stroke-width:0.264583">id </tspan></text>
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="104.99195"
y="49.168018"
id="text852-9-7-6"><tspan
sodipodi:role="line"
id="tspan850-4-8-8"
x="104.99195"
y="49.168018"
style="font-style:normal;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';stroke-width:0.264583">h(block 1)</tspan></text>
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="104.99195"
y="56.583336"
id="text852-9-7-6-8"><tspan
sodipodi:role="line"
id="tspan850-4-8-8-4"
x="104.99195"
y="56.583336"
style="font-style:normal;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';stroke-width:0.264583">h(block 2)</tspan></text>
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="104.99195"
y="64.265732"
id="text852-9-7-6-3"><tspan
sodipodi:role="line"
id="tspan850-4-8-8-1"
x="104.99195"
y="64.265732"
style="font-style:normal;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';stroke-width:0.264583">...</tspan></text>
</g>
<g
id="g898-3"
transform="translate(75.777779,38.888663)">
<rect
style="fill:none;stroke:#000000;stroke-width:0.799999;stroke-miterlimit:4;stroke-dasharray:none;stroke-dashoffset:0;stroke-opacity:1"
id="rect833-7-6"
width="47.419891"
height="29.989157"
x="95.443573"
y="24.42766" />
<rect
style="fill:none;stroke:#000000;stroke-width:0.799999;stroke-miterlimit:4;stroke-dasharray:none;stroke-dashoffset:0;stroke-opacity:1"
id="rect833-3-4-7"
width="47.419891"
height="21.608831"
x="95.443573"
y="32.807987" />
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="102.11134"
y="29.894743"
id="text852-4-5"><tspan
sodipodi:role="line"
id="tspan850-3-3"
x="102.11134"
y="29.894743"
style="font-style:normal;font-variant:normal;font-weight:bold;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';stroke-width:0.264583">Data block</tspan></text>
<path
style="fill:#ff6600;fill-opacity:1;stroke:none;stroke-width:0.0337704;stroke-opacity:1"
d="m 102.90563,41.413279 c -0.14226,0.478709 -0.31308,0.745518 -0.65239,1.018956 -0.29019,0.233843 -0.59571,0.339489 -1.03301,0.357199 -0.33087,0.0134 -0.37381,0.0074 -0.64184,-0.09013 -0.50954,-0.185343 -0.914221,-0.594142 -1.082877,-1.093901 -0.163852,-0.485526 -0.07847,-1.108159 0.210335,-1.533803 0.111448,-0.164254 0.402172,-0.444591 0.534502,-0.515415 l 0.0681,-3.63921 0.73832,-0.64992 0.88727,0.787138 -0.40985,0.445484 0.40539,0.440982 -0.40539,0.435013 0.3816,0.41947 -0.41462,0.431781 0.57335,0.585078 0.007,0.726386 c 0.13641,0.08277 0.48753,0.413601 0.59931,0.58235 0.0577,0.0871 0.13636,0.257787 0.17481,0.379304 0.0803,0.253823 0.11239,0.738377 0.0604,0.913234 z m -1.28155,0.296888 c -0.11255,-0.405265 -0.62848,-0.483569 -0.86192,-0.130817 -0.0744,0.112498 -0.088,0.164461 -0.0745,0.2847 0.0338,0.299998 0.31326,0.504498 0.58822,0.43046 0.15271,-0.04112 0.3056,-0.184909 0.34828,-0.327542 0.0392,-0.130864 0.0392,-0.115447 -1e-4,-0.256801 z"
id="path1971-0-5"
sodipodi:nodetypes="ssscsscccccccccccssscsssscc" />
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="104.99195"
y="41.008743"
id="text852-9-7-62"><tspan
sodipodi:role="line"
id="tspan850-4-8-9"
x="104.99195"
y="41.008743"
style="font-style:normal;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';stroke-width:0.264583">hash </tspan></text>
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="104.99195"
y="49.168018"
id="text852-9-7-6-1"><tspan
sodipodi:role="line"
id="tspan850-4-8-8-2"
x="104.99195"
y="49.168018"
style="font-style:normal;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';stroke-width:0.264583">data</tspan></text>
</g>
<path
style="fill:none;stroke:#000000;stroke-width:1;stroke-linecap:butt;stroke-linejoin:miter;stroke-miterlimit:4;stroke-dasharray:none;stroke-opacity:1;marker-end:url(#Arrow1Mend)"
d="M 42.105292,69.455903 89.563703,69.317144"
id="path954"
sodipodi:nodetypes="cc" />
<path
style="fill:none;stroke:#000000;stroke-width:1;stroke-linecap:butt;stroke-linejoin:miter;stroke-miterlimit:4;stroke-dasharray:none;stroke-opacity:1;marker-end:url(#marker1262)"
d="m 134.32612,77.363197 38.12618,0.260865"
id="path1258"
sodipodi:nodetypes="cc" />
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="8.6727352"
y="16.687063"
id="text852-3"><tspan
sodipodi:role="line"
id="tspan850-67"
x="8.6727352"
y="16.687063"
style="font-style:normal;font-variant:normal;font-weight:bold;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';stroke-width:0.264583">Objects table </tspan></text>
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="89.190445"
y="16.687063"
id="text852-3-5"><tspan
sodipodi:role="line"
id="tspan850-67-3"
x="89.190445"
y="16.687063"
style="font-style:normal;font-variant:normal;font-weight:bold;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';stroke-width:0.264583">Versions table </tspan></text>
<text
xml:space="preserve"
style="font-style:normal;font-weight:normal;font-size:5.64444px;line-height:1.25;font-family:sans-serif;fill:#000000;fill-opacity:1;stroke:none;stroke-width:0.264583"
x="174.55702"
y="16.687063"
id="text852-3-56"><tspan
sodipodi:role="line"
id="tspan850-67-2"
x="174.55702"
y="16.687063"
style="font-style:normal;font-variant:normal;font-weight:bold;font-stretch:normal;font-size:5.64444px;font-family:'Liberation Mono';-inkscape-font-specification:'Liberation Mono Bold';stroke-width:0.264583">Blocks table</tspan></text>
</g>
</svg>

After

Width:  |  Height:  |  Size: 30 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 52 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 37 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 97 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 199 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 145 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 13 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 174 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 38 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 14 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 87 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 81 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 124 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 84 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 81 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 81 KiB

File diff suppressed because it is too large Load Diff

After

Width:  |  Height:  |  Size: 315 KiB

File diff suppressed because one or more lines are too long

After

Width:  |  Height:  |  Size: 286 KiB

+3
View File
@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:f245bb017eb85a227d96e576ae821f165ef478d46de02535c196638aa5fc84b0
size 2575953
+480
View File
@@ -0,0 +1,480 @@
%\nonstopmode
\documentclass[aspectratio=169]{beamer}
\usepackage[utf8]{inputenc}
% \usepackage[frenchb]{babel}
\usepackage{amsmath}
\usepackage{mathtools}
\usepackage{breqn}
\usepackage{multirow}
\usetheme{boxes}
\usepackage{graphicx}
\usepackage{adjustbox}
%\useoutertheme[footline=authortitle,subsection=false]{miniframes}
%\useoutertheme[footline=authorinstitute,subsection=false]{miniframes}
\useoutertheme{infolines}
\setbeamertemplate{headline}{}
\beamertemplatenavigationsymbolsempty
\definecolor{TitleOrange}{RGB}{255,137,0}
\setbeamercolor{title}{fg=TitleOrange}
\setbeamercolor{frametitle}{fg=TitleOrange}
\definecolor{ListOrange}{RGB}{255,145,5}
\setbeamertemplate{itemize item}{\color{ListOrange}$\blacktriangleright$}
\definecolor{verygrey}{RGB}{70,70,70}
\setbeamercolor{normal text}{fg=verygrey}
\usepackage{tabu}
\usepackage{multicol}
\usepackage{vwcol}
\usepackage{stmaryrd}
\usepackage{graphicx}
\usepackage[normalem]{ulem}
\title{Introducing Garage}
\subtitle{a new storage platform for self-hosted geo-distributed clusters}
\author{Deuxfleurs Association}
\date{IMT Atlantique, 2022-06-23}
\begin{document}
\begin{frame}
\centering
\includegraphics[width=.3\linewidth]{../../sticker/Garage.pdf}
\vspace{1em}
{\large\bf Deuxfleurs Association}
\vspace{1em}
\url{https://garagehq.deuxfleurs.fr/}
Matrix channel: \texttt{\#garage:deuxfleurs.fr}
\end{frame}
\begin{frame}
\frametitle{Who we are}
\begin{columns}[t]
\begin{column}{.2\textwidth}
\centering
\adjincludegraphics[width=.4\linewidth, valign=t]{assets/alex.jpg}
\end{column}
\begin{column}{.6\textwidth}
\textbf{Alex Auvolat}\\
PhD at Inria, team WIDE; co-founder of Deuxfleurs
\end{column}
\begin{column}{.2\textwidth}
~
\end{column}
\end{columns}
\vspace{1em}
\begin{columns}[t]
\begin{column}{.2\textwidth}
~
\end{column}
\begin{column}{.6\textwidth}
\textbf{Quentin Dufour}\\
PhD at Inria, team WIDE; co-founder of Deuxfleurs
\end{column}
\begin{column}{.2\textwidth}
\centering
\adjincludegraphics[width=.5\linewidth, valign=t]{assets/quentin.jpg}
\end{column}
\end{columns}
\vspace{2em}
\begin{columns}[t]
\begin{column}{.2\textwidth}
\centering
\adjincludegraphics[width=.5\linewidth, valign=t]{assets/deuxfleurs.pdf}
\end{column}
\begin{column}{.6\textwidth}
\textbf{Deuxfleurs}\\
A non-profit self-hosting collective,\\
member of the CHATONS network
\end{column}
\begin{column}{.2\textwidth}
\centering
\adjincludegraphics[width=.7\linewidth, valign=t]{assets/logo_chatons.png}
\end{column}
\end{columns}
\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 make a \underline{stable} system}
Enterprise-grade systems typically employ:
\vspace{1em}
\begin{itemize}
\item RAID
\item Redundant power grid + UPS
\item Redundant Internet connections
\item Low-latency links
\item ...
\end{itemize}
\vspace{1em}
$\to$ it's costly and only worth it at DC scale
\end{frame}
\begin{frame}
\frametitle{How to make a \underline{resilient} system}
\only<1,4-5>{
Instead, we use:
\vspace{1em}
\begin{itemize}
\item \textcolor<2->{gray}{Commodity hardware (e.g. old desktop PCs)}
\vspace{.5em}
\item<4-> \textcolor<5->{gray}{Commodity Internet (e.g. FTTB, FTTH) and power grid}
\vspace{.5em}
\item<5-> \textcolor<6->{gray}{\textbf{Geographical redundancy} (multi-site replication)}
\end{itemize}
}
\only<2>{
\begin{center}
\includegraphics[width=.8\linewidth]{assets/atuin.jpg}
\end{center}
}
\only<3>{
\begin{center}
\includegraphics[width=.8\linewidth]{assets/neptune.jpg}
\end{center}
}
\only<6>{
\begin{center}
\includegraphics[width=.5\linewidth]{assets/inframap.jpg}
\end{center}
}
\end{frame}
\begin{frame}
\frametitle{How to make this happen}
\begin{center}
\only<1>{\includegraphics[width=.8\linewidth]{assets/slide1.png}}%
\only<2>{\includegraphics[width=.8\linewidth]{assets/slide2.png}}%
\only<3>{\includegraphics[width=.8\linewidth]{assets/slide3.png}}%
\end{center}
\end{frame}
\begin{frame}
\frametitle{Distributed file systems are slow}
File systems are complex, for example:
\vspace{1em}
\begin{itemize}
\item Concurrent modification by several processes
\vspace{1em}
\item Folder hierarchies
\vspace{1em}
\item Other requirements of the POSIX spec
\end{itemize}
\vspace{1em}
Coordination in a distributed system is costly
\vspace{1em}
Costs explode with commodity hardware / Internet connections\\
{\small (we experienced this!)}
\end{frame}
\begin{frame}
\frametitle{A simpler solution: object storage}
Only two operations:
\vspace{1em}
\begin{itemize}
\item Put an object at a key
\vspace{1em}
\item Retrieve an object from its key
\end{itemize}
\vspace{1em}
{\footnotesize (and a few others)}
\vspace{1em}
Sufficient for many applications!
\end{frame}
\begin{frame}
\frametitle{A simpler solution: object storage}
\begin{center}
\includegraphics[width=.2\linewidth]{../2020-12-02_wide-team/img/Amazon-S3.jpg}
\hspace{5em}
\includegraphics[width=.2\linewidth]{assets/minio.png}
\end{center}
\vspace{1em}
S3: a de-facto standard, many compatible applications
\vspace{1em}
MinIO is self-hostable but not suited for geo-distributed deployments
\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
\begin{center}
\includegraphics[width=.8\linewidth]{assets/garageuses.png}
\end{center}
\end{frame}
\begin{frame}
\frametitle{Overview}
\begin{center}
\only<1>{\includegraphics[width=.45\linewidth]{assets/garage2a.drawio.pdf}}%
\only<2>{\includegraphics[width=.45\linewidth]{assets/garage2b.drawio.pdf}}%
\end{center}
\end{frame}
\begin{frame}
\frametitle{Garage is \emph{location-aware}}
\begin{center}
\includegraphics[width=\linewidth]{assets/location-aware.png}
\end{center}
\vspace{2em}
Garage replicates data on different zones when possible
\end{frame}
\begin{frame}
\frametitle{Garage is \emph{location-aware}}
\begin{center}
\includegraphics[width=.8\linewidth]{assets/map.png}
\end{center}
\end{frame}
\begin{frame}
\frametitle{How to spread files over different cluster nodes?}
\textbf{Consistent hashing (DynamoDB):}
\vspace{1em}
\begin{center}
\only<1>{\includegraphics[width=.45\columnwidth]{assets/consistent_hashing_1.pdf}}%
\only<2>{\includegraphics[width=.45\columnwidth]{assets/consistent_hashing_2.pdf}}%
\only<3>{\includegraphics[width=.45\columnwidth]{assets/consistent_hashing_3.pdf}}%
\only<4>{\includegraphics[width=.45\columnwidth]{assets/consistent_hashing_4.pdf}}%
\end{center}
\end{frame}
\begin{frame}
\frametitle{How to spread files over different cluster nodes?}
\textbf{Issues with consistent hashing:}
\vspace{1em}
\begin{itemize}
\item Doesn't dispatch data based on geographical location of nodes
\vspace{1em}
\item<2-> Geographically aware adaptation, try 1:\\
data quantities not well balanced between nodes
\vspace{1em}
\item<3-> Geographically aware adaptation, try 2:\\
too many reshuffles when adding/removing nodes
\end{itemize}
\end{frame}
\begin{frame}
\frametitle{How to spread files over different cluster nodes?}
\textbf{Garage's method: build an index table}
\vspace{1em}
Realization: we can actually precompute an optimal solution
\vspace{1em}
\visible<2->{
\begin{center}
\begin{tabular}{|l|l|l|l|}
\hline
\textbf{Partition} & \textbf{Node 1} & \textbf{Node 2} & \textbf{Node 3} \\
\hline
\hline
Partition 0 & Io (jupiter) & Drosera (atuin) & Courgette (neptune) \\
\hline
Partition 1 & Datura (atuin) & Courgette (neptune) & Io (jupiter) \\
\hline
Partition 2 & Io(jupiter) & Celeri (neptune) & Drosera (atuin) \\
\hline
\hspace{1em}$\vdots$ & \hspace{1em}$\vdots$ & \hspace{1em}$\vdots$ & \hspace{1em}$\vdots$ \\
\hline
Partition 255 & Concombre (neptune) & Io (jupiter) & Drosera (atuin) \\
\hline
\end{tabular}
\end{center}
}
\vspace{1em}
\visible<3->{
The index table is built centrally using an optimal* algorithm,\\
then propagated to all nodes\\
\hfill\footnotesize *not yet optimal but will be soon
}
\end{frame}
\begin{frame}
\frametitle{Garage's internal data structures}
\centering
\includegraphics[width=.75\columnwidth]{assets/garage_tables.pdf}
\end{frame}
%\begin{frame}
% \frametitle{Garage's architecture}
% \begin{center}
% \includegraphics[width=.35\linewidth]{assets/garage.drawio.pdf}
% \end{center}
%\end{frame}
\begin{frame}
\frametitle{Garage is \emph{coordination-free}:}
\begin{itemize}
\item No Raft or Paxos
\vspace{1em}
\item Internal data types are CRDTs
\vspace{1em}
\item All nodes are equivalent (no master/leader/index node)
\end{itemize}
\vspace{2em}
$\to$ less sensitive to higher latencies between nodes
\end{frame}
\begin{frame}
\frametitle{Consistency model}
\begin{itemize}
\item Not ACID (not required by S3 spec) / not linearizable
\vspace{1em}
\item \textbf{Read-after-write consistency}\\
{\footnotesize (stronger than eventual consistency)}
\end{itemize}
\end{frame}
\begin{frame}
\frametitle{Impact on performances}
\begin{center}
\includegraphics[width=.8\linewidth]{assets/endpoint-latency-dc.png}
\end{center}
\end{frame}
\begin{frame}
\frametitle{An ever-increasing compatibility list}
\begin{center}
\includegraphics[width=.7\linewidth]{assets/compatibility.png}
\end{center}
\end{frame}
\begin{frame}
\frametitle{Further plans for Garage}
\begin{center}
\only<1>{\includegraphics[width=.8\linewidth]{assets/slideB1.png}}%
\only<2>{\includegraphics[width=.8\linewidth]{assets/slideB2.png}}%
\only<3>{\includegraphics[width=.8\linewidth]{assets/slideB3.png}}%
\end{center}
\end{frame}
\begin{frame}
\frametitle{K2V Design}
\begin{itemize}
\item A new, custom, minimal API
\vspace{1em}
\item<2-> Exposes the partitoning mechanism of Garage\\
K2V = partition key / sort key / value (like Dynamo)
\vspace{1em}
\item<3-> Coordination-free, CRDT-friendly (inspired by Riak)\\
\vspace{1em}
\item<4-> Cryptography-friendly: values are binary blobs
\end{itemize}
\end{frame}
\begin{frame}
\frametitle{Application: an e-mail storage server}
\begin{center}
\only<1>{\includegraphics[width=.9\linewidth]{assets/aerogramme.png}}%
\end{center}
\end{frame}
\begin{frame}
\frametitle{Aerogramme data model}
\begin{center}
\only<1>{\includegraphics[width=.4\linewidth]{assets/aerogramme_datatype.drawio.pdf}}%
\only<2->{\includegraphics[width=.9\linewidth]{assets/aerogramme_keys.drawio.pdf}\vspace{1em}}%
\end{center}
\visible<3->{Aerogramme encrypts all stored values for privacy\\
(Garage server administrators can't read your mail)}
\end{frame}
\begin{frame}
\frametitle{Different deployment scenarios}
\begin{center}
\only<1>{\includegraphics[width=.9\linewidth]{assets/aerogramme_components1.drawio.pdf}}%
\only<2>{\includegraphics[width=.9\linewidth]{assets/aerogramme_components2.drawio.pdf}}%
\end{center}
\end{frame}
\begin{frame}
\frametitle{A new model for building resilient software}
\begin{itemize}
\item Design a data model suited to K2V\\
{\footnotesize (see Cassandra docs on porting SQL data models to Cassandra)}
\vspace{1em}
\begin{itemize}
\item Use CRDTs or other eventually consistent data types (see e.g. Bayou)
\vspace{1em}
\item Store opaque binary blobs to provide End-to-End Encryption\\
\end{itemize}
\vspace{1em}
\item Store big blobs (files) in S3
\vspace{1em}
\item Let Garage manage sharding, replication, failover, etc.
\end{itemize}
\end{frame}
\begin{frame}
\frametitle{Research perspectives}
\begin{itemize}
\item Write about Garage's global architecture \emph{(paper in progress)}
\vspace{1em}
\item Measure and improve Garage's performances
\vspace{1em}
\item Discuss the optimal layout algorithm, provide proofs
\vspace{1em}
\item Write about our proposed architecture for (E2EE) apps over K2V+S3
\end{itemize}
\end{frame}
\begin{frame}
\frametitle{Where to find us}
\begin{center}
\includegraphics[width=.25\linewidth]{../../logo/garage_hires.png}\\
\vspace{-1em}
\url{https://garagehq.deuxfleurs.fr/}\\
\url{mailto:garagehq@deuxfleurs.fr}\\
\texttt{\#garage:deuxfleurs.fr} on Matrix
\vspace{1.5em}
\includegraphics[width=.06\linewidth]{assets/rust_logo.png}
\includegraphics[width=.13\linewidth]{assets/AGPLv3_Logo.png}
\end{center}
\end{frame}
\end{document}
%% vim: set ts=4 sw=4 tw=0 noet spelllang=en :
+1 -6
View File
@@ -156,12 +156,7 @@ impl ApiHandler for AdminApiServer {
}
Endpoint::CreateBucket => handle_create_bucket(&self.garage, req).await,
Endpoint::DeleteBucket { id } => handle_delete_bucket(&self.garage, id).await,
Endpoint::PutBucketWebsite { id } => {
handle_put_bucket_website(&self.garage, id, req).await
}
Endpoint::DeleteBucketWebsite { id } => {
handle_delete_bucket_website(&self.garage, id).await
}
Endpoint::UpdateBucket { id } => handle_update_bucket(&self.garage, id, req).await,
// Bucket-key permissions
Endpoint::BucketAllowKey => {
handle_bucket_change_key_perm(&self.garage, req, true).await
+70 -39
View File
@@ -6,7 +6,6 @@ use serde::{Deserialize, Serialize};
use garage_util::crdt::*;
use garage_util::data::*;
use garage_util::error::Error as GarageError;
use garage_util::time::*;
use garage_table::*;
@@ -15,11 +14,12 @@ use garage_model::bucket_alias_table::*;
use garage_model::bucket_table::*;
use garage_model::garage::Garage;
use garage_model::permission::*;
use garage_model::s3::object_table::*;
use crate::admin::error::*;
use crate::admin::key::ApiBucketKeyPerm;
use crate::common_error::CommonError;
use crate::helpers::parse_json_body;
use crate::helpers::{json_ok_response, parse_json_body};
pub async fn handle_list_buckets(garage: &Arc<Garage>) -> Result<Response<Body>, Error> {
let buckets = garage
@@ -60,10 +60,7 @@ pub async fn handle_list_buckets(garage: &Arc<Garage>) -> Result<Response<Body>,
})
.collect::<Vec<_>>();
let resp_json = serde_json::to_string_pretty(&res).map_err(GarageError::from)?;
Ok(Response::builder()
.status(StatusCode::OK)
.body(Body::from(resp_json))?)
Ok(json_ok_response(&res)?)
}
#[derive(Serialize)]
@@ -81,6 +78,13 @@ struct BucketLocalAlias {
alias: String,
}
#[derive(Serialize, Deserialize)]
#[serde(rename_all = "camelCase")]
struct ApiBucketQuotas {
max_size: Option<u64>,
max_objects: Option<u64>,
}
pub async fn handle_get_bucket_info(
garage: &Arc<Garage>,
id: Option<String>,
@@ -112,6 +116,14 @@ async fn bucket_info_results(
.get_existing_bucket(bucket_id)
.await?;
let counters = garage
.object_counter_table
.table
.get(&bucket_id, &EmptyKey)
.await?
.map(|x| x.filtered_values(&garage.system.ring.borrow()))
.unwrap_or_default();
let mut relevant_keys = HashMap::new();
for (k, _) in bucket
.state
@@ -152,6 +164,7 @@ async fn bucket_info_results(
let state = bucket.state.as_option().unwrap();
let quotas = state.quotas.get();
let res =
GetBucketInfoResult {
id: hex::encode(&bucket.id),
@@ -195,12 +208,19 @@ async fn bucket_info_results(
}
})
.collect::<Vec<_>>(),
objects: counters.get(OBJECTS).cloned().unwrap_or_default(),
bytes: counters.get(BYTES).cloned().unwrap_or_default(),
unfinshed_uploads: counters
.get(UNFINISHED_UPLOADS)
.cloned()
.unwrap_or_default(),
quotas: ApiBucketQuotas {
max_size: quotas.max_size,
max_objects: quotas.max_objects,
},
};
let resp_json = serde_json::to_string_pretty(&res).map_err(GarageError::from)?;
Ok(Response::builder()
.status(StatusCode::OK)
.body(Body::from(resp_json))?)
Ok(json_ok_response(&res)?)
}
#[derive(Serialize)]
@@ -212,6 +232,10 @@ struct GetBucketInfoResult {
#[serde(default)]
website_config: Option<GetBucketInfoWebsiteResult>,
keys: Vec<GetBucketInfoKey>,
objects: i64,
bytes: i64,
unfinshed_uploads: i64,
quotas: ApiBucketQuotas,
}
#[derive(Serialize)]
@@ -370,14 +394,12 @@ pub async fn handle_delete_bucket(
.body(Body::empty())?)
}
// ---- BUCKET WEBSITE CONFIGURATION ----
pub async fn handle_put_bucket_website(
pub async fn handle_update_bucket(
garage: &Arc<Garage>,
id: String,
req: Request<Body>,
) -> Result<Response<Body>, Error> {
let req = parse_json_body::<PutBucketWebsiteRequest>(req).await?;
let req = parse_json_body::<UpdateBucketRequest>(req).await?;
let bucket_id = parse_bucket_id(&id)?;
let mut bucket = garage
@@ -386,10 +408,31 @@ pub async fn handle_put_bucket_website(
.await?;
let state = bucket.state.as_option_mut().unwrap();
state.website_config.update(Some(WebsiteConfig {
index_document: req.index_document,
error_document: req.error_document,
}));
if let Some(wa) = req.website_access {
if wa.enabled {
state.website_config.update(Some(WebsiteConfig {
index_document: wa.index_document.ok_or_bad_request(
"Please specify indexDocument when enabling website access.",
)?,
error_document: wa.error_document,
}));
} else {
if wa.index_document.is_some() || wa.error_document.is_some() {
return Err(Error::bad_request(
"Cannot specify indexDocument or errorDocument when disabling website access.",
));
}
state.website_config.update(None);
}
}
if let Some(q) = req.quotas {
state.quotas.update(BucketQuotas {
max_size: q.max_size,
max_objects: q.max_objects,
});
}
garage.bucket_table.insert(&bucket).await?;
@@ -398,29 +441,17 @@ pub async fn handle_put_bucket_website(
#[derive(Deserialize)]
#[serde(rename_all = "camelCase")]
struct PutBucketWebsiteRequest {
index_document: String,
#[serde(default)]
error_document: Option<String>,
struct UpdateBucketRequest {
website_access: Option<UpdateBucketWebsiteAccess>,
quotas: Option<ApiBucketQuotas>,
}
pub async fn handle_delete_bucket_website(
garage: &Arc<Garage>,
id: String,
) -> Result<Response<Body>, Error> {
let bucket_id = parse_bucket_id(&id)?;
let mut bucket = garage
.bucket_helper()
.get_existing_bucket(bucket_id)
.await?;
let state = bucket.state.as_option_mut().unwrap();
state.website_config.update(None);
garage.bucket_table.insert(&bucket).await?;
bucket_info_results(garage, bucket_id).await
#[derive(Deserialize)]
#[serde(rename_all = "camelCase")]
struct UpdateBucketWebsiteAccess {
enabled: bool,
index_document: Option<String>,
error_document: Option<String>,
}
// ---- BUCKET/KEY PERMISSIONS ----
+7 -14
View File
@@ -7,19 +7,19 @@ use serde::{Deserialize, Serialize};
use garage_util::crdt::*;
use garage_util::data::*;
use garage_util::error::Error as GarageError;
use garage_rpc::layout::*;
use garage_model::garage::Garage;
use crate::admin::error::*;
use crate::helpers::parse_json_body;
use crate::helpers::{json_ok_response, parse_json_body};
pub async fn handle_get_cluster_status(garage: &Arc<Garage>) -> Result<Response<Body>, Error> {
let res = GetClusterStatusResponse {
node: hex::encode(garage.system.id),
garage_version: garage.system.garage_version(),
db_engine: garage.db.engine(),
known_nodes: garage
.system
.get_known_nodes()
@@ -39,10 +39,7 @@ pub async fn handle_get_cluster_status(garage: &Arc<Garage>) -> Result<Response<
layout: get_cluster_layout(garage),
};
let resp_json = serde_json::to_string_pretty(&res).map_err(GarageError::from)?;
Ok(Response::builder()
.status(StatusCode::OK)
.body(Body::from(resp_json))?)
Ok(json_ok_response(&res)?)
}
pub async fn handle_connect_cluster_nodes(
@@ -66,18 +63,13 @@ pub async fn handle_connect_cluster_nodes(
})
.collect::<Vec<_>>();
let resp_json = serde_json::to_string_pretty(&res).map_err(GarageError::from)?;
Ok(Response::builder()
.status(StatusCode::OK)
.body(Body::from(resp_json))?)
Ok(json_ok_response(&res)?)
}
pub async fn handle_get_cluster_layout(garage: &Arc<Garage>) -> Result<Response<Body>, Error> {
let res = get_cluster_layout(garage);
let resp_json = serde_json::to_string_pretty(&res).map_err(GarageError::from)?;
Ok(Response::builder()
.status(StatusCode::OK)
.body(Body::from(resp_json))?)
Ok(json_ok_response(&res)?)
}
fn get_cluster_layout(garage: &Arc<Garage>) -> GetClusterLayoutResponse {
@@ -107,6 +99,7 @@ fn get_cluster_layout(garage: &Arc<Garage>) -> GetClusterLayoutResponse {
struct GetClusterStatusResponse {
node: String,
garage_version: &'static str,
db_engine: String,
known_nodes: HashMap<String, KnownNodeResp>,
layout: GetClusterLayoutResponse,
}
+3 -2
View File
@@ -72,8 +72,9 @@ impl ApiError for Error {
}
}
fn add_http_headers(&self, _header_map: &mut HeaderMap<HeaderValue>) {
// nothing
fn add_http_headers(&self, header_map: &mut HeaderMap<HeaderValue>) {
use hyper::header;
header_map.append(header::CONTENT_TYPE, "application/json".parse().unwrap());
}
fn http_body(&self, garage_region: &str, path: &str) -> Body {
+3 -11
View File
@@ -4,15 +4,13 @@ use std::sync::Arc;
use hyper::{Body, Request, Response, StatusCode};
use serde::{Deserialize, Serialize};
use garage_util::error::Error as GarageError;
use garage_table::*;
use garage_model::garage::Garage;
use garage_model::key_table::*;
use crate::admin::error::*;
use crate::helpers::parse_json_body;
use crate::helpers::{json_ok_response, parse_json_body};
pub async fn handle_list_keys(garage: &Arc<Garage>) -> Result<Response<Body>, Error> {
let res = garage
@@ -32,10 +30,7 @@ pub async fn handle_list_keys(garage: &Arc<Garage>) -> Result<Response<Body>, Er
})
.collect::<Vec<_>>();
let resp_json = serde_json::to_string_pretty(&res).map_err(GarageError::from)?;
Ok(Response::builder()
.status(StatusCode::OK)
.body(Body::from(resp_json))?)
Ok(json_ok_response(&res)?)
}
#[derive(Serialize)]
@@ -221,10 +216,7 @@ async fn key_info_results(garage: &Arc<Garage>, key: Key) -> Result<Response<Bod
.collect::<Vec<_>>(),
};
let resp_json = serde_json::to_string_pretty(&res).map_err(GarageError::from)?;
Ok(Response::builder()
.status(StatusCode::OK)
.body(Body::from(resp_json))?)
Ok(json_ok_response(&res)?)
}
#[derive(Serialize)]
+2 -6
View File
@@ -48,10 +48,7 @@ pub enum Endpoint {
DeleteBucket {
id: String,
},
PutBucketWebsite {
id: String,
},
DeleteBucketWebsite {
UpdateBucket {
id: String,
},
// Bucket-Key Permissions
@@ -113,8 +110,7 @@ impl Endpoint {
GET "/v0/bucket" => ListBuckets,
POST "/v0/bucket" => CreateBucket,
DELETE "/v0/bucket" if id => DeleteBucket (query::id),
PUT "/v0/bucket/website" if id => PutBucketWebsite (query::id),
DELETE "/v0/bucket/website" if id => DeleteBucketWebsite (query::id),
PUT "/v0/bucket" if id => UpdateBucket (query::id),
// Bucket-key permissions
POST "/v0/bucket/allow" => BucketAllowKey,
POST "/v0/bucket/deny" => BucketDenyKey,
+1 -3
View File
@@ -150,9 +150,7 @@ impl<A: ApiHandler> ApiServer<A> {
}
Err(e) => {
let body: Body = e.http_body(&self.region, uri.path());
let mut http_error_builder = Response::builder()
.status(e.http_status_code())
.header("Content-Type", "application/xml");
let mut http_error_builder = Response::builder().status(e.http_status_code());
if let Some(header_map) = http_error_builder.headers_mut() {
e.add_http_headers(header_map)
+9 -1
View File
@@ -1,4 +1,4 @@
use hyper::{Body, Request};
use hyper::{Body, Request, Response};
use idna::domain_to_unicode;
use serde::{Deserialize, Serialize};
@@ -144,6 +144,14 @@ pub async fn parse_json_body<T: for<'de> Deserialize<'de>>(req: Request<Body>) -
Ok(resp)
}
pub fn json_ok_response<T: Serialize>(res: &T) -> Result<Response<Body>, Error> {
let resp_json = serde_json::to_string_pretty(res).map_err(garage_util::error::Error::from)?;
Ok(Response::builder()
.status(hyper::StatusCode::OK)
.header(http::header::CONTENT_TYPE, "application/json")
.body(Body::from(resp_json))?)
}
#[cfg(test)]
mod tests {
use super::*;
+3 -2
View File
@@ -110,8 +110,9 @@ impl ApiError for Error {
}
}
fn add_http_headers(&self, _header_map: &mut HeaderMap<HeaderValue>) {
// nothing
fn add_http_headers(&self, header_map: &mut HeaderMap<HeaderValue>) {
use hyper::header;
header_map.append(header::CONTENT_TYPE, "application/json".parse().unwrap());
}
fn http_body(&self, garage_region: &str, path: &str) -> Body {
+1 -1
View File
@@ -10,7 +10,7 @@ use garage_rpc::ring::Ring;
use garage_table::util::*;
use garage_model::garage::Garage;
use garage_model::k2v::counter_table::{BYTES, CONFLICTS, ENTRIES, VALUES};
use garage_model::k2v::item_table::{BYTES, CONFLICTS, ENTRIES, VALUES};
use crate::k2v::error::*;
use crate::k2v::range::read_range;
+4 -11
View File
@@ -4,7 +4,7 @@ use async_trait::async_trait;
use futures::future::Future;
use hyper::header;
use hyper::{Body, Method, Request, Response};
use hyper::{Body, Request, Response};
use opentelemetry::{trace::SpanRef, KeyValue};
@@ -167,14 +167,7 @@ impl ApiHandler for S3ApiServer {
return Err(Error::forbidden("Operation is not allowed for this key."));
}
// Look up what CORS rule might apply to response.
// Requests for methods different than GET, HEAD or POST
// are always preflighted, i.e. the browser should make
// an OPTIONS call before to check it is allowed
let matching_cors_rule = match *req.method() {
Method::GET | Method::HEAD | Method::POST => find_matching_cors_rule(&bucket, &req)?,
_ => None,
};
let matching_cors_rule = find_matching_cors_rule(&bucket, &req)?;
let resp = match endpoint {
Endpoint::HeadObject {
@@ -219,7 +212,7 @@ impl ApiHandler for S3ApiServer {
.await
}
Endpoint::PutObject { key } => {
handle_put(garage, req, bucket_id, &key, content_sha256).await
handle_put(garage, req, &bucket, &key, content_sha256).await
}
Endpoint::AbortMultipartUpload { key, upload_id } => {
handle_abort_multipart_upload(garage, bucket_id, &key, &upload_id).await
@@ -233,7 +226,7 @@ impl ApiHandler for S3ApiServer {
garage,
req,
&bucket_name,
bucket_id,
&bucket,
&key,
&upload_id,
content_sha256,
+3
View File
@@ -172,6 +172,9 @@ impl ApiError for Error {
fn add_http_headers(&self, header_map: &mut HeaderMap<HeaderValue>) {
use hyper::header;
header_map.append(header::CONTENT_TYPE, "application/xml".parse().unwrap());
#[allow(clippy::single_match)]
match self {
Error::InvalidRange((_, len)) => {
+10 -5
View File
@@ -22,7 +22,7 @@ use crate::signature::payload::{parse_date, verify_v4};
pub async fn handle_post_object(
garage: Arc<Garage>,
req: Request<Body>,
bucket: String,
bucket_name: String,
) -> Result<Response<Body>, Error> {
let boundary = req
.headers()
@@ -126,13 +126,18 @@ pub async fn handle_post_object(
let bucket_id = garage
.bucket_helper()
.resolve_bucket(&bucket, &api_key)
.resolve_bucket(&bucket_name, &api_key)
.await?;
if !api_key.allow_write(&bucket_id) {
return Err(Error::forbidden("Operation is not allowed for this key."));
}
let bucket = garage
.bucket_helper()
.get_existing_bucket(bucket_id)
.await?;
let decoded_policy = base64::decode(&policy).ok_or_bad_request("Invalid policy")?;
let decoded_policy: Policy =
serde_json::from_slice(&decoded_policy).ok_or_bad_request("Invalid policy")?;
@@ -227,7 +232,7 @@ pub async fn handle_post_object(
garage,
headers,
StreamLimiter::new(stream, conditions.content_length),
bucket_id,
&bucket,
&key,
None,
None,
@@ -244,7 +249,7 @@ pub async fn handle_post_object(
{
target
.query_pairs_mut()
.append_pair("bucket", &bucket)
.append_pair("bucket", &bucket_name)
.append_pair("key", &key)
.append_pair("etag", &etag);
let target = target.to_string();
@@ -289,7 +294,7 @@ pub async fn handle_post_object(
let xml = s3_xml::PostObject {
xmlns: (),
location: s3_xml::Value(location),
bucket: s3_xml::Value(bucket),
bucket: s3_xml::Value(bucket_name),
key: s3_xml::Value(key),
etag: s3_xml::Value(etag),
};
+103 -26
View File
@@ -1,4 +1,4 @@
use std::collections::{BTreeMap, BTreeSet, VecDeque};
use std::collections::{BTreeMap, BTreeSet, HashMap, VecDeque};
use std::sync::Arc;
use futures::prelude::*;
@@ -14,7 +14,9 @@ use garage_util::error::Error as GarageError;
use garage_util::time::*;
use garage_block::manager::INLINE_THRESHOLD;
use garage_model::bucket_table::Bucket;
use garage_model::garage::Garage;
use garage_model::index_counter::CountedItem;
use garage_model::s3::block_ref_table::*;
use garage_model::s3::object_table::*;
use garage_model::s3::version_table::*;
@@ -26,7 +28,7 @@ use crate::signature::verify_signed_content;
pub async fn handle_put(
garage: Arc<Garage>,
req: Request<Body>,
bucket_id: Uuid,
bucket: &Bucket,
key: &str,
content_sha256: Option<Hash>,
) -> Result<Response<Body>, Error> {
@@ -46,7 +48,7 @@ pub async fn handle_put(
garage,
headers,
body,
bucket_id,
bucket,
key,
content_md5,
content_sha256,
@@ -59,7 +61,7 @@ pub(crate) async fn save_stream<S: Stream<Item = Result<Bytes, Error>> + Unpin>(
garage: Arc<Garage>,
headers: ObjectVersionHeaders,
body: S,
bucket_id: Uuid,
bucket: &Bucket,
key: &str,
content_md5: Option<String>,
content_sha256: Option<FixedBytes32>,
@@ -80,6 +82,7 @@ pub(crate) async fn save_stream<S: Stream<Item = Result<Bytes, Error>> + Unpin>(
let data_md5sum_hex = hex::encode(data_md5sum);
let data_sha256sum = sha256sum(&first_block[..]);
let size = first_block.len() as u64;
ensure_checksum_matches(
data_md5sum.as_slice(),
@@ -88,20 +91,22 @@ pub(crate) async fn save_stream<S: Stream<Item = Result<Bytes, Error>> + Unpin>(
content_sha256,
)?;
check_quotas(&garage, bucket, key, size).await?;
let object_version = ObjectVersion {
uuid: version_uuid,
timestamp: version_timestamp,
state: ObjectVersionState::Complete(ObjectVersionData::Inline(
ObjectVersionMeta {
headers,
size: first_block.len() as u64,
size,
etag: data_md5sum_hex.clone(),
},
first_block,
)),
};
let object = Object::new(bucket_id, key.into(), vec![object_version]);
let object = Object::new(bucket.id, key.into(), vec![object_version]);
garage.object_table.insert(&object).await?;
return Ok((version_uuid, data_md5sum_hex));
@@ -114,36 +119,42 @@ pub(crate) async fn save_stream<S: Stream<Item = Result<Bytes, Error>> + Unpin>(
timestamp: version_timestamp,
state: ObjectVersionState::Uploading(headers.clone()),
};
let object = Object::new(bucket_id, key.into(), vec![object_version.clone()]);
let object = Object::new(bucket.id, key.into(), vec![object_version.clone()]);
garage.object_table.insert(&object).await?;
// Initialize corresponding entry in version table
// Write this entry now, even with empty block list,
// to prevent block_ref entries from being deleted (they can be deleted
// if the reference a version that isn't found in the version table)
let version = Version::new(version_uuid, bucket_id, key.into(), false);
let version = Version::new(version_uuid, bucket.id, key.into(), false);
garage.version_table.insert(&version).await?;
// Transfer data and verify checksum
let first_block_hash = blake2sum(&first_block[..]);
let tx_result = read_and_put_blocks(
&garage,
&version,
1,
first_block,
first_block_hash,
&mut chunker,
)
.await
.and_then(|(total_size, data_md5sum, data_sha256sum)| {
let tx_result = (|| async {
let (total_size, data_md5sum, data_sha256sum) = read_and_put_blocks(
&garage,
&version,
1,
first_block,
first_block_hash,
&mut chunker,
)
.await?;
ensure_checksum_matches(
data_md5sum.as_slice(),
data_sha256sum,
content_md5.as_deref(),
content_sha256,
)
.map(|()| (total_size, data_md5sum))
});
)?;
check_quotas(&garage, bucket, key, total_size).await?;
Ok((total_size, data_md5sum))
})()
.await;
// If something went wrong, clean up
let (total_size, md5sum_arr) = match tx_result {
@@ -151,7 +162,7 @@ pub(crate) async fn save_stream<S: Stream<Item = Result<Bytes, Error>> + Unpin>(
Err(e) => {
// Mark object as aborted, this will free the blocks further down
object_version.state = ObjectVersionState::Aborted;
let object = Object::new(bucket_id, key.into(), vec![object_version.clone()]);
let object = Object::new(bucket.id, key.into(), vec![object_version.clone()]);
garage.object_table.insert(&object).await?;
return Err(e);
}
@@ -167,7 +178,7 @@ pub(crate) async fn save_stream<S: Stream<Item = Result<Bytes, Error>> + Unpin>(
},
first_block_hash,
));
let object = Object::new(bucket_id, key.into(), vec![object_version]);
let object = Object::new(bucket.id, key.into(), vec![object_version]);
garage.object_table.insert(&object).await?;
Ok((version_uuid, md5sum_hex))
@@ -200,6 +211,64 @@ fn ensure_checksum_matches(
Ok(())
}
/// Check that inserting this object with this size doesn't exceed bucket quotas
async fn check_quotas(
garage: &Arc<Garage>,
bucket: &Bucket,
key: &str,
size: u64,
) -> Result<(), Error> {
let quotas = bucket.state.as_option().unwrap().quotas.get();
if quotas.max_objects.is_none() && quotas.max_size.is_none() {
return Ok(());
};
let key = key.to_string();
let (prev_object, counters) = futures::try_join!(
garage.object_table.get(&bucket.id, &key),
garage.object_counter_table.table.get(&bucket.id, &EmptyKey),
)?;
let counters = counters
.map(|x| x.filtered_values(&garage.system.ring.borrow()))
.unwrap_or_default();
let (prev_cnt_obj, prev_cnt_size) = match prev_object {
Some(o) => {
let prev_cnt = o.counts().into_iter().collect::<HashMap<_, _>>();
(
prev_cnt.get(OBJECTS).cloned().unwrap_or_default(),
prev_cnt.get(BYTES).cloned().unwrap_or_default(),
)
}
None => (0, 0),
};
let cnt_obj_diff = 1 - prev_cnt_obj;
let cnt_size_diff = size as i64 - prev_cnt_size;
if let Some(mo) = quotas.max_objects {
let current_objects = counters.get(OBJECTS).cloned().unwrap_or_default();
if cnt_obj_diff > 0 && current_objects + cnt_obj_diff > mo as i64 {
return Err(Error::forbidden(format!(
"Object quota is reached, maximum objects for this bucket: {}",
mo
)));
}
}
if let Some(ms) = quotas.max_size {
let current_size = counters.get(BYTES).cloned().unwrap_or_default();
if cnt_size_diff > 0 && current_size + cnt_size_diff > ms as i64 {
return Err(Error::forbidden(format!(
"Bucket size quota is reached, maximum total size of objects for this bucket: {}. The bucket is already {} bytes, and this object would add {} bytes.",
ms, current_size, size
)));
}
}
Ok(())
}
async fn read_and_put_blocks<S: Stream<Item = Result<Bytes, Error>> + Unpin>(
garage: &Garage,
version: &Version,
@@ -473,7 +542,7 @@ pub async fn handle_complete_multipart_upload(
garage: Arc<Garage>,
req: Request<Body>,
bucket_name: &str,
bucket_id: Uuid,
bucket: &Bucket,
key: &str,
upload_id: &str,
content_sha256: Option<Hash>,
@@ -497,7 +566,7 @@ pub async fn handle_complete_multipart_upload(
// Get object and version
let key = key.to_string();
let (object, version) = futures::try_join!(
garage.object_table.get(&bucket_id, &key),
garage.object_table.get(&bucket.id, &key),
garage.version_table.get(&version_uuid, &EmptyKey),
)?;
@@ -590,6 +659,14 @@ pub async fn handle_complete_multipart_upload(
// Calculate total size of final object
let total_size = version.blocks.items().iter().map(|x| x.1.size).sum();
if let Err(e) = check_quotas(&garage, bucket, &key, total_size).await {
object_version.state = ObjectVersionState::Aborted;
let final_object = Object::new(bucket.id, key.clone(), vec![object_version]);
garage.object_table.insert(&final_object).await?;
return Err(e);
}
// Write final object version
object_version.state = ObjectVersionState::Complete(ObjectVersionData::FirstBlock(
ObjectVersionMeta {
@@ -600,7 +677,7 @@ pub async fn handle_complete_multipart_upload(
version.blocks.items()[0].1.hash,
));
let final_object = Object::new(bucket_id, key.clone(), vec![object_version]);
let final_object = Object::new(bucket.id, key.clone(), vec![object_version]);
garage.object_table.insert(&final_object).await?;
// Send response saying ok we're done
+2 -2
View File
@@ -14,12 +14,14 @@ path = "lib.rs"
# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
[dependencies]
garage_db = { version = "0.8.0", path = "../db" }
garage_rpc = { version = "0.7.0", path = "../rpc" }
garage_util = { version = "0.7.0", path = "../util" }
garage_table = { version = "0.7.0", path = "../table" }
opentelemetry = "0.17"
arc-swap = "1.5"
async-trait = "0.1.7"
bytes = "1.0"
hex = "0.4"
@@ -27,8 +29,6 @@ tracing = "0.1.30"
rand = "0.8"
zstd = { version = "0.9", default-features = false }
sled = "0.34"
rmp-serde = "0.15"
serde = { version = "1.0", default-features = false, features = ["derive", "rc"] }
serde_bytes = "0.11"
+1
View File
@@ -2,6 +2,7 @@
extern crate tracing;
pub mod manager;
pub mod repair;
mod block;
mod metrics;
+168 -204
View File
@@ -1,26 +1,30 @@
use std::convert::TryInto;
use std::path::{Path, PathBuf};
use std::path::PathBuf;
use std::sync::Arc;
use std::time::Duration;
use arc_swap::ArcSwapOption;
use async_trait::async_trait;
use serde::{Deserialize, Serialize};
use futures::future::*;
use futures::select;
use tokio::fs;
use tokio::io::{AsyncReadExt, AsyncWriteExt};
use tokio::sync::{watch, Mutex, Notify};
use tokio::select;
use tokio::sync::{mpsc, watch, Mutex, Notify};
use opentelemetry::{
trace::{FutureExt as OtelFutureExt, TraceContextExt, Tracer},
Context, KeyValue,
};
use garage_db as db;
use garage_db::counted_tree_hack::CountedTree;
use garage_util::background::*;
use garage_util::data::*;
use garage_util::error::*;
use garage_util::metrics::RecordDuration;
use garage_util::sled_counter::SledCountedTree;
use garage_util::time::*;
use garage_util::tranquilizer::Tranquilizer;
@@ -32,6 +36,7 @@ use garage_table::replication::{TableReplication, TableShardedReplication};
use crate::block::*;
use crate::metrics::*;
use crate::rc::*;
use crate::repair::*;
/// Size under which data will be stored inlined in database instead of as files
pub const INLINE_THRESHOLD: usize = 3072;
@@ -89,16 +94,18 @@ pub struct BlockManager {
mutation_lock: Mutex<BlockManagerLocked>,
rc: BlockRc,
pub(crate) rc: BlockRc,
resync_queue: SledCountedTree,
resync_queue: CountedTree,
resync_notify: Notify,
resync_errors: SledCountedTree,
resync_errors: CountedTree,
system: Arc<System>,
pub(crate) system: Arc<System>,
endpoint: Arc<Endpoint<BlockRpc, Self>>,
metrics: BlockManagerMetrics,
tx_scrub_command: ArcSwapOption<mpsc::Sender<ScrubWorkerCommand>>,
}
// This custom struct contains functions that must only be ran
@@ -106,9 +113,15 @@ pub struct BlockManager {
// it INSIDE a Mutex.
struct BlockManagerLocked();
enum ResyncIterResult {
BusyDidSomething,
BusyDidNothing,
IdleFor(Duration),
}
impl BlockManager {
pub fn new(
db: &sled::Db,
db: &db::Db,
data_dir: PathBuf,
compression_level: Option<i32>,
background_tranquility: u32,
@@ -123,12 +136,14 @@ impl BlockManager {
let resync_queue = db
.open_tree("block_local_resync_queue")
.expect("Unable to open block_local_resync_queue tree");
let resync_queue = SledCountedTree::new(resync_queue);
let resync_queue =
CountedTree::new(resync_queue).expect("Could not count block_local_resync_queue");
let resync_errors = db
.open_tree("block_local_resync_errors")
.expect("Unable to open block_local_resync_errors tree");
let resync_errors = SledCountedTree::new(resync_errors);
let resync_errors =
CountedTree::new(resync_errors).expect("Could not count block_local_resync_errors");
let endpoint = system
.netapp
@@ -151,10 +166,11 @@ impl BlockManager {
system,
endpoint,
metrics,
tx_scrub_command: ArcSwapOption::new(None),
});
block_manager.endpoint.set_handler(block_manager.clone());
block_manager.clone().spawn_background_worker();
block_manager.clone().spawn_background_workers();
block_manager
}
@@ -212,98 +228,81 @@ impl BlockManager {
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.
for (i, entry) in self.rc.rc.iter().enumerate() {
let (hash, _) = entry?;
let hash = Hash::try_from(&hash[..]).unwrap();
self.put_to_resync(&hash, Duration::from_secs(0))?;
if i & 0xFF == 0 && *must_exit.borrow() {
return Ok(());
}
}
// 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))
.map_err(Into::into)
},
must_exit,
)
.await
}
/// 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
}
/// Get lenght of resync queue
pub fn resync_queue_len(&self) -> usize {
self.resync_queue.len()
pub fn resync_queue_len(&self) -> Result<usize, Error> {
// This currently can't return an error because the CountedTree hack
// doesn't error on .len(), but this will change when we remove the hack
// (hopefully someday!)
Ok(self.resync_queue.len())
}
/// Get number of blocks that have an error
pub fn resync_errors_len(&self) -> usize {
self.resync_errors.len()
pub fn resync_errors_len(&self) -> Result<usize, Error> {
// (see resync_queue_len comment)
Ok(self.resync_errors.len())
}
/// Get number of items in the refcount table
pub fn rc_len(&self) -> usize {
self.rc.rc.len()
pub fn rc_len(&self) -> Result<usize, Error> {
Ok(self.rc.rc.len()?)
}
/// Send command to start/stop/manager scrub worker
pub async fn send_scrub_command(&self, cmd: ScrubWorkerCommand) {
let _ = self
.tx_scrub_command
.load()
.as_ref()
.unwrap()
.send(cmd)
.await;
}
//// ----- 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> {
if self.rc.block_incref(hash)? {
pub fn block_incref(
self: &Arc<Self>,
tx: &mut db::Transaction,
hash: Hash,
) -> db::TxOpResult<()> {
if self.rc.block_incref(tx, &hash)? {
// When the reference counter is incremented, there is
// normally a node that is responsible for sending us the
// data of the block. However that operation may fail,
// so in all cases we add the block here to the todo list
// to check later that it arrived correctly, and if not
// we will fecth it from someone.
self.put_to_resync(hash, 2 * BLOCK_RW_TIMEOUT)?;
let this = self.clone();
tokio::spawn(async move {
if let Err(e) = this.put_to_resync(&hash, 2 * BLOCK_RW_TIMEOUT) {
error!("Block {:?} could not be put in resync queue: {}.", hash, e);
}
});
}
Ok(())
}
/// Decrement the number of time a block is used
pub fn block_decref(&self, hash: &Hash) -> Result<(), Error> {
if self.rc.block_decref(hash)? {
pub fn block_decref(
self: &Arc<Self>,
tx: &mut db::Transaction,
hash: Hash,
) -> db::TxOpResult<()> {
if self.rc.block_decref(tx, &hash)? {
// When the RC is decremented, it might drop to zero,
// indicating that we don't need the block.
// There is a delay before we garbage collect it;
// make sure that it is handled in the resync loop
// after that delay has passed.
self.put_to_resync(hash, BLOCK_GC_DELAY + Duration::from_secs(10))?;
let this = self.clone();
tokio::spawn(async move {
if let Err(e) = this.put_to_resync(&hash, BLOCK_GC_DELAY + Duration::from_secs(10))
{
error!("Block {:?} could not be put in resync queue: {}.", hash, e);
}
});
}
Ok(())
}
@@ -328,7 +327,7 @@ impl BlockManager {
}
/// Read block from disk, verifying it's integrity
async fn read_block(&self, hash: &Hash) -> Result<BlockRpc, Error> {
pub(crate) async fn read_block(&self, hash: &Hash) -> Result<BlockRpc, Error> {
let data = self
.read_block_internal(hash)
.bound_record_duration(&self.metrics.block_read_duration)
@@ -492,23 +491,28 @@ impl BlockManager {
// for times that are earlier than the exponential back-off delay
// is a natural condition that is handled properly).
fn spawn_background_worker(self: Arc<Self>) {
fn spawn_background_workers(self: Arc<Self>) {
// Launch a background workers for background resync loop processing
let background = self.system.background.clone();
let worker = ResyncWorker::new(self.clone());
tokio::spawn(async move {
tokio::time::sleep(Duration::from_secs(10)).await;
background.spawn_worker("block resync worker".into(), move |must_exit| {
self.resync_loop(must_exit)
});
background.spawn_worker(worker);
});
// Launch a background worker for data store scrubs
let (scrub_tx, scrub_rx) = mpsc::channel(1);
self.tx_scrub_command.store(Some(Arc::new(scrub_tx)));
let scrub_worker = ScrubWorker::new(self.clone(), scrub_rx);
self.system.background.spawn_worker(scrub_worker);
}
fn put_to_resync(&self, hash: &Hash, delay: Duration) -> Result<(), sled::Error> {
pub(crate) fn put_to_resync(&self, hash: &Hash, delay: Duration) -> db::Result<()> {
let when = now_msec() + delay.as_millis() as u64;
self.put_to_resync_at(hash, when)
}
fn put_to_resync_at(&self, hash: &Hash, when: u64) -> Result<(), sled::Error> {
fn put_to_resync_at(&self, hash: &Hash, when: u64) -> db::Result<()> {
trace!("Put resync_queue: {} {:?}", when, hash);
let mut key = u64::to_be_bytes(when).to_vec();
key.extend(hash.as_ref());
@@ -517,43 +521,8 @@ impl BlockManager {
Ok(())
}
async fn resync_loop(self: Arc<Self>, mut must_exit: watch::Receiver<bool>) {
let mut tranquilizer = Tranquilizer::new(30);
while !*must_exit.borrow() {
match self.resync_iter(&mut must_exit).await {
Ok(true) => {
tranquilizer.tranquilize(self.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();
}
}
}
}
// The result of resync_iter is:
// - Ok(true) -> a block was processed (successfully or not)
// - Ok(false) -> no block was processed, but we are ready for the next iteration
// - Err(_) -> a Sled error occurred when reading/writing from resync_queue/resync_errors
async fn resync_iter(
&self,
must_exit: &mut watch::Receiver<bool>,
) -> Result<bool, sled::Error> {
if let Some(first_pair_res) = self.resync_queue.iter().next() {
let (time_bytes, hash_bytes) = first_pair_res?;
async fn resync_iter(&self) -> Result<ResyncIterResult, db::Error> {
if let Some((time_bytes, hash_bytes)) = self.resync_queue.first()? {
let time_msec = u64::from_be_bytes(time_bytes[0..8].try_into().unwrap());
let now = now_msec();
@@ -561,7 +530,7 @@ impl BlockManager {
let hash = Hash::try_from(&hash_bytes[..]).unwrap();
if let Some(ec) = self.resync_errors.get(hash.as_slice())? {
let ec = ErrorCounter::decode(ec);
let ec = ErrorCounter::decode(&ec);
if now < ec.next_try() {
// if next retry after an error is not yet,
// don't do resync and return early, but still
@@ -572,7 +541,7 @@ impl BlockManager {
// (we want to do the remove after the insert to ensure
// that the item is not lost if we crash in-between)
self.resync_queue.remove(time_bytes)?;
return Ok(false);
return Ok(ResyncIterResult::BusyDidNothing);
}
}
@@ -602,7 +571,7 @@ impl BlockManager {
warn!("Error when resyncing {:?}: {}", hash, e);
let err_counter = match self.resync_errors.get(hash.as_slice())? {
Some(ec) => ErrorCounter::decode(ec).add1(now + 1),
Some(ec) => ErrorCounter::decode(&ec).add1(now + 1),
None => ErrorCounter::new(now + 1),
};
@@ -619,15 +588,11 @@ impl BlockManager {
self.resync_queue.remove(time_bytes)?;
}
Ok(true)
Ok(ResyncIterResult::BusyDidSomething)
} else {
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)
Ok(ResyncIterResult::IdleFor(Duration::from_millis(
time_msec - now,
)))
}
} else {
// Here we wait either for a notification that an item has been
@@ -636,13 +601,7 @@ impl BlockManager {
// between the time we checked the queue and the first poll
// to resync_notify.notified(): if that happens, we'll just loop
// back 10 seconds later, which is fine.
let delay = tokio::time::sleep(Duration::from_secs(10));
select! {
_ = delay.fuse() => {},
_ = self.resync_notify.notified().fuse() => {},
_ = must_exit.changed().fuse() => {},
}
Ok(false)
Ok(ResyncIterResult::IdleFor(Duration::from_secs(10)))
}
}
@@ -757,72 +716,6 @@ impl BlockManager {
Ok(())
}
// ---- Utility: iteration on files in the data directory ----
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(|_| ())
}
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<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?;
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 = if let Ok(n) = name.into_string() {
n
} else {
continue;
};
let ent_type = data_dir_ent.file_type().await?;
let name = name.strip_suffix(".zst").unwrap_or(&name);
if name.len() == 2 && hex::decode(&name).is_ok() && ent_type.is_dir() {
state = self
.for_each_file_rec(&data_dir_ent.path(), state, f, must_exit)
.await?;
} else if name.len() == 64 {
let hash_bytes = if let Ok(h) = hex::decode(&name) {
h
} else {
continue;
};
let mut hash = [0u8; 32];
hash.copy_from_slice(&hash_bytes[..]);
state = f(state, hash.into()).await?;
}
}
Ok(state)
}
.boxed()
}
}
#[async_trait]
@@ -841,6 +734,77 @@ impl EndpointHandler<BlockRpc> for BlockManager {
}
}
struct ResyncWorker {
manager: Arc<BlockManager>,
tranquilizer: Tranquilizer,
next_delay: Duration,
}
impl ResyncWorker {
fn new(manager: Arc<BlockManager>) -> Self {
Self {
manager,
tranquilizer: Tranquilizer::new(30),
next_delay: Duration::from_secs(10),
}
}
}
#[async_trait]
impl Worker for ResyncWorker {
fn name(&self) -> String {
"Block resync worker".into()
}
fn info(&self) -> Option<String> {
let mut ret = vec![];
let qlen = self.manager.resync_queue_len().unwrap_or(0);
let elen = self.manager.resync_errors_len().unwrap_or(0);
if qlen > 0 {
ret.push(format!("{} blocks in queue", qlen));
}
if elen > 0 {
ret.push(format!("{} blocks in error state", elen));
}
if !ret.is_empty() {
Some(ret.join(", "))
} else {
None
}
}
async fn work(&mut self, _must_exit: &mut watch::Receiver<bool>) -> Result<WorkerState, Error> {
self.tranquilizer.reset();
match self.manager.resync_iter().await {
Ok(ResyncIterResult::BusyDidSomething) => Ok(self
.tranquilizer
.tranquilize_worker(self.manager.background_tranquility)),
Ok(ResyncIterResult::BusyDidNothing) => Ok(WorkerState::Busy),
Ok(ResyncIterResult::IdleFor(delay)) => {
self.next_delay = delay;
Ok(WorkerState::Idle)
}
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?)
// Here we just give the error to the worker manager,
// it will print it to the logs and increment a counter
Err(e.into())
}
}
}
async fn wait_for_work(&mut self, _must_exit: &watch::Receiver<bool>) -> WorkerState {
select! {
_ = tokio::time::sleep(self.next_delay) => (),
_ = self.manager.resync_notify.notified() => (),
};
WorkerState::Busy
}
}
struct BlockStatus {
exists: bool,
needed: RcEntry,
@@ -966,7 +930,7 @@ impl ErrorCounter {
}
}
fn decode(data: sled::IVec) -> Self {
fn decode(data: &[u8]) -> Self {
Self {
errors: u64::from_be_bytes(data[0..8].try_into().unwrap()),
last_try: u64::from_be_bytes(data[8..16].try_into().unwrap()),
+2 -2
View File
@@ -1,6 +1,6 @@
use opentelemetry::{global, metrics::*};
use garage_util::sled_counter::SledCountedTree;
use garage_db::counted_tree_hack::CountedTree;
/// TableMetrics reference all counter used for metrics
pub struct BlockManagerMetrics {
@@ -23,7 +23,7 @@ pub struct BlockManagerMetrics {
}
impl BlockManagerMetrics {
pub fn new(resync_queue: SledCountedTree, resync_errors: SledCountedTree) -> Self {
pub fn new(resync_queue: CountedTree, resync_errors: CountedTree) -> Self {
let meter = global::meter("garage_model/block");
Self {
_resync_queue_len: meter
+32 -17
View File
@@ -1,5 +1,7 @@
use std::convert::TryInto;
use garage_db as db;
use garage_util::data::*;
use garage_util::error::*;
use garage_util::time::*;
@@ -7,31 +9,41 @@ use garage_util::time::*;
use crate::manager::BLOCK_GC_DELAY;
pub struct BlockRc {
pub(crate) rc: sled::Tree,
pub(crate) rc: db::Tree,
}
impl BlockRc {
pub(crate) fn new(rc: sled::Tree) -> Self {
pub(crate) fn new(rc: db::Tree) -> Self {
Self { rc }
}
/// Increment the reference counter associated to a hash.
/// Returns true if the RC goes from zero to nonzero.
pub(crate) fn block_incref(&self, hash: &Hash) -> Result<bool, Error> {
let old_rc = self
.rc
.fetch_and_update(&hash, |old| RcEntry::parse_opt(old).increment().serialize())?;
let old_rc = RcEntry::parse_opt(old_rc);
pub(crate) fn block_incref(
&self,
tx: &mut db::Transaction,
hash: &Hash,
) -> db::TxOpResult<bool> {
let old_rc = RcEntry::parse_opt(tx.get(&self.rc, &hash)?);
match old_rc.increment().serialize() {
Some(x) => tx.insert(&self.rc, &hash, x)?,
None => unreachable!(),
};
Ok(old_rc.is_zero())
}
/// Decrement the reference counter associated to a hash.
/// Returns true if the RC is now zero.
pub(crate) fn block_decref(&self, hash: &Hash) -> Result<bool, Error> {
let new_rc = self
.rc
.update_and_fetch(&hash, |old| RcEntry::parse_opt(old).decrement().serialize())?;
let new_rc = RcEntry::parse_opt(new_rc);
pub(crate) fn block_decref(
&self,
tx: &mut db::Transaction,
hash: &Hash,
) -> db::TxOpResult<bool> {
let new_rc = RcEntry::parse_opt(tx.get(&self.rc, &hash)?).decrement();
match new_rc.serialize() {
Some(x) => tx.insert(&self.rc, &hash, x)?,
None => tx.remove(&self.rc, &hash)?,
};
Ok(matches!(new_rc, RcEntry::Deletable { .. }))
}
@@ -44,12 +56,15 @@ impl BlockRc {
/// deletion time has passed
pub(crate) fn clear_deleted_block_rc(&self, hash: &Hash) -> Result<(), Error> {
let now = now_msec();
self.rc.update_and_fetch(&hash, |rcval| {
let updated = match RcEntry::parse_opt(rcval) {
RcEntry::Deletable { at_time } if now > at_time => RcEntry::Absent,
v => v,
self.rc.db().transaction(|mut tx| {
let rcval = RcEntry::parse_opt(tx.get(&self.rc, &hash)?);
match rcval {
RcEntry::Deletable { at_time } if now > at_time => {
tx.remove(&self.rc, &hash)?;
}
_ => (),
};
updated.serialize()
tx.commit(())
})?;
Ok(())
}
+444
View File
@@ -0,0 +1,444 @@
use core::ops::Bound;
use std::path::PathBuf;
use std::sync::Arc;
use std::time::Duration;
use async_trait::async_trait;
use serde::{Deserialize, Serialize};
use tokio::fs;
use tokio::select;
use tokio::sync::mpsc;
use tokio::sync::watch;
use garage_util::background::*;
use garage_util::data::*;
use garage_util::error::*;
use garage_util::persister::Persister;
use garage_util::time::*;
use garage_util::tranquilizer::Tranquilizer;
use crate::manager::*;
const SCRUB_INTERVAL: Duration = Duration::from_secs(3600 * 24 * 30); // full scrub every 30 days
pub struct RepairWorker {
manager: Arc<BlockManager>,
next_start: Option<Hash>,
block_iter: Option<BlockStoreIterator>,
}
impl RepairWorker {
pub fn new(manager: Arc<BlockManager>) -> Self {
Self {
manager,
next_start: None,
block_iter: None,
}
}
}
#[async_trait]
impl Worker for RepairWorker {
fn name(&self) -> String {
"Block repair worker".into()
}
fn info(&self) -> Option<String> {
match self.block_iter.as_ref() {
None => {
let idx_bytes = self
.next_start
.as_ref()
.map(|x| x.as_slice())
.unwrap_or(&[]);
let idx_bytes = if idx_bytes.len() > 4 {
&idx_bytes[..4]
} else {
idx_bytes
};
Some(format!("Phase 1: {}", hex::encode(idx_bytes)))
}
Some(bi) => Some(format!("Phase 2: {:.2}% done", bi.progress() * 100.)),
}
}
async fn work(&mut self, _must_exit: &mut watch::Receiver<bool>) -> Result<WorkerState, Error> {
match self.block_iter.as_mut() {
None => {
// Phase 1: Repair blocks from RC table.
// We have to do this complicated two-step process where we first read a bunch
// of hashes from the RC table, and then insert them in the to-resync queue,
// because of SQLite. Basically, as long as we have an iterator on a DB table,
// we can't do anything else on the DB. The naive approach (which we had previously)
// of just iterating on the RC table and inserting items one to one in the resync
// queue can't work here, it would just provoke a deadlock in the SQLite adapter code.
// This is mostly because the Rust bindings for SQLite assume a worst-case scenario
// where SQLite is not compiled in thread-safe mode, so we have to wrap everything
// in a mutex (see db/sqlite_adapter.rs and discussion in PR #322).
// TODO: maybe do this with tokio::task::spawn_blocking ?
let mut batch_of_hashes = vec![];
let start_bound = match self.next_start.as_ref() {
None => Bound::Unbounded,
Some(x) => Bound::Excluded(x.as_slice()),
};
for entry in self
.manager
.rc
.rc
.range::<&[u8], _>((start_bound, Bound::Unbounded))?
{
let (hash, _) = entry?;
let hash = Hash::try_from(&hash[..]).unwrap();
batch_of_hashes.push(hash);
if batch_of_hashes.len() >= 1000 {
break;
}
}
if batch_of_hashes.is_empty() {
// move on to phase 2
self.block_iter = Some(BlockStoreIterator::new(&self.manager));
return Ok(WorkerState::Busy);
}
for hash in batch_of_hashes.into_iter() {
self.manager.put_to_resync(&hash, Duration::from_secs(0))?;
self.next_start = Some(hash)
}
Ok(WorkerState::Busy)
}
Some(bi) => {
// Phase 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.
if let Some(hash) = bi.next().await? {
self.manager.put_to_resync(&hash, Duration::from_secs(0))?;
Ok(WorkerState::Busy)
} else {
Ok(WorkerState::Done)
}
}
}
}
async fn wait_for_work(&mut self, _must_exit: &watch::Receiver<bool>) -> WorkerState {
unreachable!()
}
}
// ----
pub struct ScrubWorker {
manager: Arc<BlockManager>,
rx_cmd: mpsc::Receiver<ScrubWorkerCommand>,
work: ScrubWorkerState,
tranquilizer: Tranquilizer,
persister: Persister<ScrubWorkerPersisted>,
persisted: ScrubWorkerPersisted,
}
#[derive(Serialize, Deserialize)]
struct ScrubWorkerPersisted {
tranquility: u32,
time_last_complete_scrub: u64,
corruptions_detected: u64,
}
enum ScrubWorkerState {
Running(BlockStoreIterator),
Paused(BlockStoreIterator, u64), // u64 = time when to resume scrub
Finished,
}
impl Default for ScrubWorkerState {
fn default() -> Self {
ScrubWorkerState::Finished
}
}
#[derive(Debug)]
pub enum ScrubWorkerCommand {
Start,
Pause(Duration),
Resume,
Cancel,
SetTranquility(u32),
}
impl ScrubWorker {
pub fn new(manager: Arc<BlockManager>, rx_cmd: mpsc::Receiver<ScrubWorkerCommand>) -> Self {
let persister = Persister::new(&manager.system.metadata_dir, "scrub_info");
let persisted = match persister.load() {
Ok(v) => v,
Err(_) => ScrubWorkerPersisted {
time_last_complete_scrub: 0,
tranquility: 4,
corruptions_detected: 0,
},
};
Self {
manager,
rx_cmd,
work: ScrubWorkerState::Finished,
tranquilizer: Tranquilizer::new(30),
persister,
persisted,
}
}
async fn handle_cmd(&mut self, cmd: ScrubWorkerCommand) {
match cmd {
ScrubWorkerCommand::Start => {
self.work = match std::mem::take(&mut self.work) {
ScrubWorkerState::Finished => {
let iterator = BlockStoreIterator::new(&self.manager);
ScrubWorkerState::Running(iterator)
}
work => {
error!("Cannot start scrub worker: already running!");
work
}
};
}
ScrubWorkerCommand::Pause(dur) => {
self.work = match std::mem::take(&mut self.work) {
ScrubWorkerState::Running(it) | ScrubWorkerState::Paused(it, _) => {
ScrubWorkerState::Paused(it, now_msec() + dur.as_millis() as u64)
}
work => {
error!("Cannot pause scrub worker: not running!");
work
}
};
}
ScrubWorkerCommand::Resume => {
self.work = match std::mem::take(&mut self.work) {
ScrubWorkerState::Paused(it, _) => ScrubWorkerState::Running(it),
work => {
error!("Cannot resume scrub worker: not paused!");
work
}
};
}
ScrubWorkerCommand::Cancel => {
self.work = match std::mem::take(&mut self.work) {
ScrubWorkerState::Running(_) | ScrubWorkerState::Paused(_, _) => {
ScrubWorkerState::Finished
}
work => {
error!("Cannot cancel scrub worker: not running!");
work
}
}
}
ScrubWorkerCommand::SetTranquility(t) => {
self.persisted.tranquility = t;
if let Err(e) = self.persister.save_async(&self.persisted).await {
error!("Could not save new tranquilitiy value: {}", e);
}
}
}
}
}
#[async_trait]
impl Worker for ScrubWorker {
fn name(&self) -> String {
"Block scrub worker".into()
}
fn info(&self) -> Option<String> {
let s = match &self.work {
ScrubWorkerState::Running(bsi) => format!(
"{:.2}% done (tranquility = {})",
bsi.progress() * 100.,
self.persisted.tranquility
),
ScrubWorkerState::Paused(bsi, rt) => {
format!(
"Paused, {:.2}% done, resumes at {}",
bsi.progress() * 100.,
msec_to_rfc3339(*rt)
)
}
ScrubWorkerState::Finished => format!(
"Last completed scrub: {}",
msec_to_rfc3339(self.persisted.time_last_complete_scrub)
),
};
Some(format!(
"{} ; corruptions detected: {}",
s, self.persisted.corruptions_detected
))
}
async fn work(&mut self, _must_exit: &mut watch::Receiver<bool>) -> Result<WorkerState, Error> {
match self.rx_cmd.try_recv() {
Ok(cmd) => self.handle_cmd(cmd).await,
Err(mpsc::error::TryRecvError::Disconnected) => return Ok(WorkerState::Done),
Err(mpsc::error::TryRecvError::Empty) => (),
};
match &mut self.work {
ScrubWorkerState::Running(bsi) => {
self.tranquilizer.reset();
if let Some(hash) = bsi.next().await? {
match self.manager.read_block(&hash).await {
Err(Error::CorruptData(_)) => {
error!("Found corrupt data block during scrub: {:?}", hash);
self.persisted.corruptions_detected += 1;
self.persister.save_async(&self.persisted).await?;
}
Err(e) => return Err(e),
_ => (),
};
Ok(self
.tranquilizer
.tranquilize_worker(self.persisted.tranquility))
} else {
self.persisted.time_last_complete_scrub = now_msec();
self.persister.save_async(&self.persisted).await?;
self.work = ScrubWorkerState::Finished;
self.tranquilizer.clear();
Ok(WorkerState::Idle)
}
}
_ => Ok(WorkerState::Idle),
}
}
async fn wait_for_work(&mut self, _must_exit: &watch::Receiver<bool>) -> WorkerState {
let (wait_until, command) = match &self.work {
ScrubWorkerState::Running(_) => return WorkerState::Busy,
ScrubWorkerState::Paused(_, resume_time) => (*resume_time, ScrubWorkerCommand::Resume),
ScrubWorkerState::Finished => (
self.persisted.time_last_complete_scrub + SCRUB_INTERVAL.as_millis() as u64,
ScrubWorkerCommand::Start,
),
};
let now = now_msec();
if now >= wait_until {
self.handle_cmd(command).await;
return WorkerState::Busy;
}
let delay = Duration::from_millis(wait_until - now);
select! {
_ = tokio::time::sleep(delay) => self.handle_cmd(command).await,
cmd = self.rx_cmd.recv() => if let Some(cmd) = cmd {
self.handle_cmd(cmd).await;
} else {
return WorkerState::Done;
}
}
match &self.work {
ScrubWorkerState::Running(_) => WorkerState::Busy,
_ => WorkerState::Idle,
}
}
}
// ----
struct BlockStoreIterator {
path: Vec<ReadingDir>,
}
enum ReadingDir {
Pending(PathBuf),
Read {
subpaths: Vec<fs::DirEntry>,
pos: usize,
},
}
impl BlockStoreIterator {
fn new(manager: &BlockManager) -> Self {
let root_dir = manager.data_dir.clone();
Self {
path: vec![ReadingDir::Pending(root_dir)],
}
}
/// Returns progress done, between 0 and 1
fn progress(&self) -> f32 {
if self.path.is_empty() {
1.0
} else {
let mut ret = 0.0;
let mut next_div = 1;
for p in self.path.iter() {
match p {
ReadingDir::Pending(_) => break,
ReadingDir::Read { subpaths, pos } => {
next_div *= subpaths.len();
ret += ((*pos - 1) as f32) / (next_div as f32);
}
}
}
ret
}
}
async fn next(&mut self) -> Result<Option<Hash>, Error> {
loop {
let last_path = match self.path.last_mut() {
None => return Ok(None),
Some(lp) => lp,
};
if let ReadingDir::Pending(path) = last_path {
let mut reader = fs::read_dir(&path).await?;
let mut subpaths = vec![];
while let Some(ent) = reader.next_entry().await? {
subpaths.push(ent);
}
*last_path = ReadingDir::Read { subpaths, pos: 0 };
}
let (subpaths, pos) = match *last_path {
ReadingDir::Read {
ref subpaths,
ref mut pos,
} => (subpaths, pos),
ReadingDir::Pending(_) => unreachable!(),
};
let data_dir_ent = match subpaths.get(*pos) {
None => {
self.path.pop();
continue;
}
Some(ent) => {
*pos += 1;
ent
}
};
let name = data_dir_ent.file_name();
let name = if let Ok(n) = name.into_string() {
n
} else {
continue;
};
let ent_type = data_dir_ent.file_type().await?;
let name = name.strip_suffix(".zst").unwrap_or(&name);
if name.len() == 2 && hex::decode(&name).is_ok() && ent_type.is_dir() {
let path = data_dir_ent.path();
self.path.push(ReadingDir::Pending(path));
} else if name.len() == 64 {
if let Ok(h) = hex::decode(&name) {
let mut hash = [0u8; 32];
hash.copy_from_slice(&h);
return Ok(Some(hash.into()));
}
}
}
}
}
+36
View File
@@ -0,0 +1,36 @@
[package]
name = "garage_db"
version = "0.8.0"
authors = ["Alex Auvolat <alex@adnab.me>"]
edition = "2018"
license = "AGPL-3.0"
description = "Abstraction over multiple key/value storage engines that supports transactions"
repository = "https://git.deuxfleurs.fr/Deuxfleurs/garage"
readme = "../../README.md"
[lib]
path = "lib.rs"
[[bin]]
name = "convert"
path = "bin/convert.rs"
required-features = ["cli"]
[dependencies]
err-derive = "0.3"
hexdump = "0.1"
tracing = "0.1.30"
heed = "0.11"
rusqlite = { version = "0.27", features = ["bundled"] }
sled = "0.34"
# cli deps
clap = { version = "3.1.18", optional = true, features = ["derive", "env"] }
pretty_env_logger = { version = "0.4", optional = true }
[dev-dependencies]
mktemp = "0.4"
[features]
cli = ["clap", "pretty_env_logger"]
+69
View File
@@ -0,0 +1,69 @@
use std::path::PathBuf;
use garage_db::*;
use clap::Parser;
/// K2V command line interface
#[derive(Parser, Debug)]
#[clap(author, version, about, long_about = None)]
struct Args {
/// Input DB path
#[clap(short = 'i')]
input_path: PathBuf,
/// Input DB engine
#[clap(short = 'a')]
input_engine: String,
/// Output DB path
#[clap(short = 'o')]
output_path: PathBuf,
/// Output DB engine
#[clap(short = 'b')]
output_engine: String,
}
fn main() {
let args = Args::parse();
pretty_env_logger::init();
match do_conversion(args) {
Ok(()) => println!("Success!"),
Err(e) => eprintln!("Error: {}", e),
}
}
fn do_conversion(args: Args) -> Result<()> {
let input = open_db(args.input_path, args.input_engine)?;
let output = open_db(args.output_path, args.output_engine)?;
output.import(&input)?;
Ok(())
}
fn open_db(path: PathBuf, engine: String) -> Result<Db> {
match engine.as_str() {
"sled" => {
let db = sled_adapter::sled::Config::default().path(&path).open()?;
Ok(sled_adapter::SledDb::init(db))
}
"sqlite" | "sqlite3" | "rusqlite" => {
let db = sqlite_adapter::rusqlite::Connection::open(&path)?;
Ok(sqlite_adapter::SqliteDb::init(db))
}
"lmdb" | "heed" => {
std::fs::create_dir_all(&path).map_err(|e| {
Error(format!("Unable to create LMDB data directory: {}", e).into())
})?;
let map_size = lmdb_adapter::recommended_map_size();
let db = lmdb_adapter::heed::EnvOpenOptions::new()
.max_dbs(100)
.map_size(map_size)
.open(&path)
.unwrap();
Ok(lmdb_adapter::LmdbDb::init(db))
}
e => Err(Error(format!("Invalid DB engine: {}", e).into())),
}
}
+127
View File
@@ -0,0 +1,127 @@
//! This hack allows a db tree to keep in RAM a counter of the number of entries
//! it contains, which is used to call .len() on it. This is usefull only for
//! the sled backend where .len() otherwise would have to traverse the whole
//! tree to count items. For sqlite and lmdb, this is mostly useless (but
//! hopefully not harmfull!). Note that a CountedTree cannot be part of a
//! transaction.
use std::sync::{
atomic::{AtomicUsize, Ordering},
Arc,
};
use crate::{Result, Tree, TxError, Value, ValueIter};
#[derive(Clone)]
pub struct CountedTree(Arc<CountedTreeInternal>);
struct CountedTreeInternal {
tree: Tree,
len: AtomicUsize,
}
impl CountedTree {
pub fn new(tree: Tree) -> Result<Self> {
let len = tree.len()?;
Ok(Self(Arc::new(CountedTreeInternal {
tree,
len: AtomicUsize::new(len),
})))
}
pub fn len(&self) -> usize {
self.0.len.load(Ordering::SeqCst)
}
pub fn is_empty(&self) -> bool {
self.len() == 0
}
pub fn get<K: AsRef<[u8]>>(&self, key: K) -> Result<Option<Value>> {
self.0.tree.get(key)
}
pub fn first(&self) -> Result<Option<(Value, Value)>> {
self.0.tree.first()
}
pub fn iter(&self) -> Result<ValueIter<'_>> {
self.0.tree.iter()
}
// ---- writing functions ----
pub fn insert<K, V>(&self, key: K, value: V) -> Result<Option<Value>>
where
K: AsRef<[u8]>,
V: AsRef<[u8]>,
{
let old_val = self.0.tree.insert(key, value)?;
if old_val.is_none() {
self.0.len.fetch_add(1, Ordering::SeqCst);
}
Ok(old_val)
}
pub fn remove<K: AsRef<[u8]>>(&self, key: K) -> Result<Option<Value>> {
let old_val = self.0.tree.remove(key)?;
if old_val.is_some() {
self.0.len.fetch_sub(1, Ordering::SeqCst);
}
Ok(old_val)
}
pub fn compare_and_swap<K, OV, NV>(
&self,
key: K,
expected_old: Option<OV>,
new: Option<NV>,
) -> Result<bool>
where
K: AsRef<[u8]>,
OV: AsRef<[u8]>,
NV: AsRef<[u8]>,
{
let old_some = expected_old.is_some();
let new_some = new.is_some();
let tx_res = self.0.tree.db().transaction(|mut tx| {
let old_val = tx.get(&self.0.tree, &key)?;
let is_same = match (&old_val, &expected_old) {
(None, None) => true,
(Some(x), Some(y)) if x == y.as_ref() => true,
_ => false,
};
if is_same {
match &new {
Some(v) => {
tx.insert(&self.0.tree, &key, v)?;
}
None => {
tx.remove(&self.0.tree, &key)?;
}
}
tx.commit(())
} else {
tx.abort(())
}
});
match tx_res {
Ok(()) => {
match (old_some, new_some) {
(false, true) => {
self.0.len.fetch_add(1, Ordering::SeqCst);
}
(true, false) => {
self.0.len.fetch_sub(1, Ordering::SeqCst);
}
_ => (),
}
Ok(true)
}
Err(TxError::Abort(())) => Ok(false),
Err(TxError::Db(e)) => Err(e),
}
}
}
+409
View File
@@ -0,0 +1,409 @@
#[macro_use]
extern crate tracing;
pub mod lmdb_adapter;
pub mod sled_adapter;
pub mod sqlite_adapter;
pub mod counted_tree_hack;
#[cfg(test)]
pub mod test;
use core::ops::{Bound, RangeBounds};
use std::borrow::Cow;
use std::cell::Cell;
use std::sync::Arc;
use err_derive::Error;
#[derive(Clone)]
pub struct Db(pub(crate) Arc<dyn IDb>);
pub struct Transaction<'a>(&'a mut dyn ITx);
#[derive(Clone)]
pub struct Tree(Arc<dyn IDb>, usize);
pub type Value = Vec<u8>;
pub type ValueIter<'a> = Box<dyn std::iter::Iterator<Item = Result<(Value, Value)>> + 'a>;
pub type TxValueIter<'a> = Box<dyn std::iter::Iterator<Item = TxOpResult<(Value, Value)>> + 'a>;
// ----
#[derive(Debug, Error)]
#[error(display = "{}", _0)]
pub struct Error(pub Cow<'static, str>);
pub type Result<T> = std::result::Result<T, Error>;
#[derive(Debug, Error)]
#[error(display = "{}", _0)]
pub struct TxOpError(pub(crate) Error);
pub type TxOpResult<T> = std::result::Result<T, TxOpError>;
pub enum TxError<E> {
Abort(E),
Db(Error),
}
pub type TxResult<R, E> = std::result::Result<R, TxError<E>>;
impl<E> From<TxOpError> for TxError<E> {
fn from(e: TxOpError) -> TxError<E> {
TxError::Db(e.0)
}
}
pub fn unabort<R, E>(res: TxResult<R, E>) -> TxOpResult<std::result::Result<R, E>> {
match res {
Ok(v) => Ok(Ok(v)),
Err(TxError::Abort(e)) => Ok(Err(e)),
Err(TxError::Db(e)) => Err(TxOpError(e)),
}
}
// ----
impl Db {
pub fn engine(&self) -> String {
self.0.engine()
}
pub fn open_tree<S: AsRef<str>>(&self, name: S) -> Result<Tree> {
let tree_id = self.0.open_tree(name.as_ref())?;
Ok(Tree(self.0.clone(), tree_id))
}
pub fn list_trees(&self) -> Result<Vec<String>> {
self.0.list_trees()
}
pub fn transaction<R, E, F>(&self, fun: F) -> TxResult<R, E>
where
F: Fn(Transaction<'_>) -> TxResult<R, E>,
{
let f = TxFn {
function: fun,
result: Cell::new(None),
};
let tx_res = self.0.transaction(&f);
let ret = f
.result
.into_inner()
.expect("Transaction did not store result");
match tx_res {
Ok(()) => {
assert!(matches!(ret, Ok(_)));
ret
}
Err(TxError::Abort(())) => {
assert!(matches!(ret, Err(TxError::Abort(_))));
ret
}
Err(TxError::Db(e2)) => match ret {
// Ok was stored -> the error occured when finalizing
// transaction
Ok(_) => Err(TxError::Db(e2)),
// An error was already stored: that's the one we want to
// return
Err(TxError::Db(e)) => Err(TxError::Db(e)),
_ => unreachable!(),
},
}
}
pub fn import(&self, other: &Db) -> Result<()> {
let existing_trees = self.list_trees()?;
if !existing_trees.is_empty() {
return Err(Error(
format!(
"destination database already contains data: {:?}",
existing_trees
)
.into(),
));
}
let tree_names = other.list_trees()?;
for name in tree_names {
let tree = self.open_tree(&name)?;
if tree.len()? > 0 {
return Err(Error(format!("tree {} already contains data", name).into()));
}
let ex_tree = other.open_tree(&name)?;
let tx_res = self.transaction(|mut tx| {
let mut i = 0;
for item in ex_tree.iter().map_err(TxError::Abort)? {
let (k, v) = item.map_err(TxError::Abort)?;
tx.insert(&tree, k, v)?;
i += 1;
if i % 1000 == 0 {
println!("{}: imported {}", name, i);
}
}
tx.commit(i)
});
let total = match tx_res {
Err(TxError::Db(e)) => return Err(e),
Err(TxError::Abort(e)) => return Err(e),
Ok(x) => x,
};
println!("{}: finished importing, {} items", name, total);
}
Ok(())
}
}
#[allow(clippy::len_without_is_empty)]
impl Tree {
#[inline]
pub fn db(&self) -> Db {
Db(self.0.clone())
}
#[inline]
pub fn get<T: AsRef<[u8]>>(&self, key: T) -> Result<Option<Value>> {
self.0.get(self.1, key.as_ref())
}
#[inline]
pub fn len(&self) -> Result<usize> {
self.0.len(self.1)
}
#[inline]
pub fn first(&self) -> Result<Option<(Value, Value)>> {
self.iter()?.next().transpose()
}
#[inline]
pub fn get_gt<T: AsRef<[u8]>>(&self, from: T) -> Result<Option<(Value, Value)>> {
self.range((Bound::Excluded(from), Bound::Unbounded))?
.next()
.transpose()
}
/// Returns the old value if there was one
#[inline]
pub fn insert<T: AsRef<[u8]>, U: AsRef<[u8]>>(
&self,
key: T,
value: U,
) -> Result<Option<Value>> {
self.0.insert(self.1, key.as_ref(), value.as_ref())
}
/// Returns the old value if there was one
#[inline]
pub fn remove<T: AsRef<[u8]>>(&self, key: T) -> Result<Option<Value>> {
self.0.remove(self.1, key.as_ref())
}
/// Clears all values from the tree
#[inline]
pub fn clear(&self) -> Result<()> {
self.0.clear(self.1)
}
#[inline]
pub fn iter(&self) -> Result<ValueIter<'_>> {
self.0.iter(self.1)
}
#[inline]
pub fn iter_rev(&self) -> Result<ValueIter<'_>> {
self.0.iter_rev(self.1)
}
#[inline]
pub fn range<K, R>(&self, range: R) -> Result<ValueIter<'_>>
where
K: AsRef<[u8]>,
R: RangeBounds<K>,
{
let sb = range.start_bound();
let eb = range.end_bound();
self.0.range(self.1, get_bound(sb), get_bound(eb))
}
#[inline]
pub fn range_rev<K, R>(&self, range: R) -> Result<ValueIter<'_>>
where
K: AsRef<[u8]>,
R: RangeBounds<K>,
{
let sb = range.start_bound();
let eb = range.end_bound();
self.0.range_rev(self.1, get_bound(sb), get_bound(eb))
}
}
#[allow(clippy::len_without_is_empty)]
impl<'a> Transaction<'a> {
#[inline]
pub fn get<T: AsRef<[u8]>>(&self, tree: &Tree, key: T) -> TxOpResult<Option<Value>> {
self.0.get(tree.1, key.as_ref())
}
#[inline]
pub fn len(&self, tree: &Tree) -> TxOpResult<usize> {
self.0.len(tree.1)
}
/// Returns the old value if there was one
#[inline]
pub fn insert<T: AsRef<[u8]>, U: AsRef<[u8]>>(
&mut self,
tree: &Tree,
key: T,
value: U,
) -> TxOpResult<Option<Value>> {
self.0.insert(tree.1, key.as_ref(), value.as_ref())
}
/// Returns the old value if there was one
#[inline]
pub fn remove<T: AsRef<[u8]>>(&mut self, tree: &Tree, key: T) -> TxOpResult<Option<Value>> {
self.0.remove(tree.1, key.as_ref())
}
#[inline]
pub fn iter(&self, tree: &Tree) -> TxOpResult<TxValueIter<'_>> {
self.0.iter(tree.1)
}
#[inline]
pub fn iter_rev(&self, tree: &Tree) -> TxOpResult<TxValueIter<'_>> {
self.0.iter_rev(tree.1)
}
#[inline]
pub fn range<K, R>(&self, tree: &Tree, range: R) -> TxOpResult<TxValueIter<'_>>
where
K: AsRef<[u8]>,
R: RangeBounds<K>,
{
let sb = range.start_bound();
let eb = range.end_bound();
self.0.range(tree.1, get_bound(sb), get_bound(eb))
}
#[inline]
pub fn range_rev<K, R>(&self, tree: &Tree, range: R) -> TxOpResult<TxValueIter<'_>>
where
K: AsRef<[u8]>,
R: RangeBounds<K>,
{
let sb = range.start_bound();
let eb = range.end_bound();
self.0.range_rev(tree.1, get_bound(sb), get_bound(eb))
}
// ----
#[inline]
pub fn abort<R, E>(self, e: E) -> TxResult<R, E> {
Err(TxError::Abort(e))
}
#[inline]
pub fn commit<R, E>(self, r: R) -> TxResult<R, E> {
Ok(r)
}
}
// ---- Internal interfaces
pub(crate) trait IDb: Send + Sync {
fn engine(&self) -> String;
fn open_tree(&self, name: &str) -> Result<usize>;
fn list_trees(&self) -> Result<Vec<String>>;
fn get(&self, tree: usize, key: &[u8]) -> Result<Option<Value>>;
fn len(&self, tree: usize) -> Result<usize>;
fn insert(&self, tree: usize, key: &[u8], value: &[u8]) -> Result<Option<Value>>;
fn remove(&self, tree: usize, key: &[u8]) -> Result<Option<Value>>;
fn clear(&self, tree: usize) -> Result<()>;
fn iter(&self, tree: usize) -> Result<ValueIter<'_>>;
fn iter_rev(&self, tree: usize) -> Result<ValueIter<'_>>;
fn range<'r>(
&self,
tree: usize,
low: Bound<&'r [u8]>,
high: Bound<&'r [u8]>,
) -> Result<ValueIter<'_>>;
fn range_rev<'r>(
&self,
tree: usize,
low: Bound<&'r [u8]>,
high: Bound<&'r [u8]>,
) -> Result<ValueIter<'_>>;
fn transaction(&self, f: &dyn ITxFn) -> TxResult<(), ()>;
}
pub(crate) trait ITx {
fn get(&self, tree: usize, key: &[u8]) -> TxOpResult<Option<Value>>;
fn len(&self, tree: usize) -> TxOpResult<usize>;
fn insert(&mut self, tree: usize, key: &[u8], value: &[u8]) -> TxOpResult<Option<Value>>;
fn remove(&mut self, tree: usize, key: &[u8]) -> TxOpResult<Option<Value>>;
fn iter(&self, tree: usize) -> TxOpResult<TxValueIter<'_>>;
fn iter_rev(&self, tree: usize) -> TxOpResult<TxValueIter<'_>>;
fn range<'r>(
&self,
tree: usize,
low: Bound<&'r [u8]>,
high: Bound<&'r [u8]>,
) -> TxOpResult<TxValueIter<'_>>;
fn range_rev<'r>(
&self,
tree: usize,
low: Bound<&'r [u8]>,
high: Bound<&'r [u8]>,
) -> TxOpResult<TxValueIter<'_>>;
}
pub(crate) trait ITxFn {
fn try_on(&self, tx: &mut dyn ITx) -> TxFnResult;
}
pub(crate) enum TxFnResult {
Ok,
Abort,
DbErr,
}
struct TxFn<F, R, E>
where
F: Fn(Transaction<'_>) -> TxResult<R, E>,
{
function: F,
result: Cell<Option<TxResult<R, E>>>,
}
impl<F, R, E> ITxFn for TxFn<F, R, E>
where
F: Fn(Transaction<'_>) -> TxResult<R, E>,
{
fn try_on(&self, tx: &mut dyn ITx) -> TxFnResult {
let res = (self.function)(Transaction(tx));
let res2 = match &res {
Ok(_) => TxFnResult::Ok,
Err(TxError::Abort(_)) => TxFnResult::Abort,
Err(TxError::Db(_)) => TxFnResult::DbErr,
};
self.result.set(Some(res));
res2
}
}
// ----
fn get_bound<K: AsRef<[u8]>>(b: Bound<&K>) -> Bound<&[u8]> {
match b {
Bound::Included(v) => Bound::Included(v.as_ref()),
Bound::Excluded(v) => Bound::Excluded(v.as_ref()),
Bound::Unbounded => Bound::Unbounded,
}
}
+350
View File
@@ -0,0 +1,350 @@
use core::ops::Bound;
use core::ptr::NonNull;
use std::collections::HashMap;
use std::convert::TryInto;
use std::sync::{Arc, RwLock};
use heed::types::ByteSlice;
use heed::{BytesDecode, Env, RoTxn, RwTxn, UntypedDatabase as Database};
use crate::{
Db, Error, IDb, ITx, ITxFn, Result, TxError, TxFnResult, TxOpError, TxOpResult, TxResult,
TxValueIter, Value, ValueIter,
};
pub use heed;
// -- err
impl From<heed::Error> for Error {
fn from(e: heed::Error) -> Error {
Error(format!("LMDB: {}", e).into())
}
}
impl From<heed::Error> for TxOpError {
fn from(e: heed::Error) -> TxOpError {
TxOpError(e.into())
}
}
// -- db
pub struct LmdbDb {
db: heed::Env,
trees: RwLock<(Vec<Database>, HashMap<String, usize>)>,
}
impl LmdbDb {
pub fn init(db: Env) -> Db {
let s = Self {
db,
trees: RwLock::new((Vec::new(), HashMap::new())),
};
Db(Arc::new(s))
}
fn get_tree(&self, i: usize) -> Result<Database> {
self.trees
.read()
.unwrap()
.0
.get(i)
.cloned()
.ok_or_else(|| Error("invalid tree id".into()))
}
}
impl IDb for LmdbDb {
fn engine(&self) -> String {
"LMDB (using Heed crate)".into()
}
fn open_tree(&self, name: &str) -> Result<usize> {
let mut trees = self.trees.write().unwrap();
if let Some(i) = trees.1.get(name) {
Ok(*i)
} else {
let tree = self.db.create_database(Some(name))?;
let i = trees.0.len();
trees.0.push(tree);
trees.1.insert(name.to_string(), i);
Ok(i)
}
}
fn list_trees(&self) -> Result<Vec<String>> {
let tree0 = match self.db.open_database::<heed::types::Str, ByteSlice>(None)? {
Some(x) => x,
None => return Ok(vec![]),
};
let mut ret = vec![];
let tx = self.db.read_txn()?;
for item in tree0.iter(&tx)? {
let (tree_name, _) = item?;
ret.push(tree_name.to_string());
}
drop(tx);
let mut ret2 = vec![];
for tree_name in ret {
if self
.db
.open_database::<ByteSlice, ByteSlice>(Some(&tree_name))?
.is_some()
{
ret2.push(tree_name);
}
}
Ok(ret2)
}
// ----
fn get(&self, tree: usize, key: &[u8]) -> Result<Option<Value>> {
let tree = self.get_tree(tree)?;
let tx = self.db.read_txn()?;
let val = tree.get(&tx, key)?;
match val {
None => Ok(None),
Some(v) => Ok(Some(v.to_vec())),
}
}
fn len(&self, tree: usize) -> Result<usize> {
let tree = self.get_tree(tree)?;
let tx = self.db.read_txn()?;
Ok(tree.len(&tx)?.try_into().unwrap())
}
fn insert(&self, tree: usize, key: &[u8], value: &[u8]) -> Result<Option<Value>> {
let tree = self.get_tree(tree)?;
let mut tx = self.db.write_txn()?;
let old_val = tree.get(&tx, key)?.map(Vec::from);
tree.put(&mut tx, key, value)?;
tx.commit()?;
Ok(old_val)
}
fn remove(&self, tree: usize, key: &[u8]) -> Result<Option<Value>> {
let tree = self.get_tree(tree)?;
let mut tx = self.db.write_txn()?;
let old_val = tree.get(&tx, key)?.map(Vec::from);
tree.delete(&mut tx, key)?;
tx.commit()?;
Ok(old_val)
}
fn clear(&self, tree: usize) -> Result<()> {
let tree = self.get_tree(tree)?;
let mut tx = self.db.write_txn()?;
tree.clear(&mut tx)?;
tx.commit()?;
Ok(())
}
fn iter(&self, tree: usize) -> Result<ValueIter<'_>> {
let tree = self.get_tree(tree)?;
let tx = self.db.read_txn()?;
TxAndIterator::make(tx, |tx| Ok(tree.iter(tx)?))
}
fn iter_rev(&self, tree: usize) -> Result<ValueIter<'_>> {
let tree = self.get_tree(tree)?;
let tx = self.db.read_txn()?;
TxAndIterator::make(tx, |tx| Ok(tree.rev_iter(tx)?))
}
fn range<'r>(
&self,
tree: usize,
low: Bound<&'r [u8]>,
high: Bound<&'r [u8]>,
) -> Result<ValueIter<'_>> {
let tree = self.get_tree(tree)?;
let tx = self.db.read_txn()?;
TxAndIterator::make(tx, |tx| Ok(tree.range(tx, &(low, high))?))
}
fn range_rev<'r>(
&self,
tree: usize,
low: Bound<&'r [u8]>,
high: Bound<&'r [u8]>,
) -> Result<ValueIter<'_>> {
let tree = self.get_tree(tree)?;
let tx = self.db.read_txn()?;
TxAndIterator::make(tx, |tx| Ok(tree.rev_range(tx, &(low, high))?))
}
// ----
fn transaction(&self, f: &dyn ITxFn) -> TxResult<(), ()> {
let trees = self.trees.read().unwrap();
let mut tx = LmdbTx {
trees: &trees.0[..],
tx: self
.db
.write_txn()
.map_err(Error::from)
.map_err(TxError::Db)?,
};
let res = f.try_on(&mut tx);
match res {
TxFnResult::Ok => {
tx.tx.commit().map_err(Error::from).map_err(TxError::Db)?;
Ok(())
}
TxFnResult::Abort => {
tx.tx.abort().map_err(Error::from).map_err(TxError::Db)?;
Err(TxError::Abort(()))
}
TxFnResult::DbErr => {
tx.tx.abort().map_err(Error::from).map_err(TxError::Db)?;
Err(TxError::Db(Error(
"(this message will be discarded)".into(),
)))
}
}
}
}
// ----
struct LmdbTx<'a> {
trees: &'a [Database],
tx: RwTxn<'a, 'a>,
}
impl<'a> LmdbTx<'a> {
fn get_tree(&self, i: usize) -> TxOpResult<&Database> {
self.trees.get(i).ok_or_else(|| {
TxOpError(Error(
"invalid tree id (it might have been openned after the transaction started)".into(),
))
})
}
}
impl<'a> ITx for LmdbTx<'a> {
fn get(&self, tree: usize, key: &[u8]) -> TxOpResult<Option<Value>> {
let tree = self.get_tree(tree)?;
match tree.get(&self.tx, key)? {
Some(v) => Ok(Some(v.to_vec())),
None => Ok(None),
}
}
fn len(&self, _tree: usize) -> TxOpResult<usize> {
unimplemented!(".len() in transaction not supported with LMDB backend")
}
fn insert(&mut self, tree: usize, key: &[u8], value: &[u8]) -> TxOpResult<Option<Value>> {
let tree = *self.get_tree(tree)?;
let old_val = tree.get(&self.tx, key)?.map(Vec::from);
tree.put(&mut self.tx, key, value)?;
Ok(old_val)
}
fn remove(&mut self, tree: usize, key: &[u8]) -> TxOpResult<Option<Value>> {
let tree = *self.get_tree(tree)?;
let old_val = tree.get(&self.tx, key)?.map(Vec::from);
tree.delete(&mut self.tx, key)?;
Ok(old_val)
}
fn iter(&self, _tree: usize) -> TxOpResult<TxValueIter<'_>> {
unimplemented!("Iterators in transactions not supported with LMDB backend");
}
fn iter_rev(&self, _tree: usize) -> TxOpResult<TxValueIter<'_>> {
unimplemented!("Iterators in transactions not supported with LMDB backend");
}
fn range<'r>(
&self,
_tree: usize,
_low: Bound<&'r [u8]>,
_high: Bound<&'r [u8]>,
) -> TxOpResult<TxValueIter<'_>> {
unimplemented!("Iterators in transactions not supported with LMDB backend");
}
fn range_rev<'r>(
&self,
_tree: usize,
_low: Bound<&'r [u8]>,
_high: Bound<&'r [u8]>,
) -> TxOpResult<TxValueIter<'_>> {
unimplemented!("Iterators in transactions not supported with LMDB backend");
}
}
// ----
type IteratorItem<'a> = heed::Result<(
<ByteSlice as BytesDecode<'a>>::DItem,
<ByteSlice as BytesDecode<'a>>::DItem,
)>;
struct TxAndIterator<'a, I>
where
I: Iterator<Item = IteratorItem<'a>> + 'a,
{
tx: RoTxn<'a>,
iter: Option<I>,
}
impl<'a, I> TxAndIterator<'a, I>
where
I: Iterator<Item = IteratorItem<'a>> + 'a,
{
fn make<F>(tx: RoTxn<'a>, iterfun: F) -> Result<ValueIter<'a>>
where
F: FnOnce(&'a RoTxn<'a>) -> Result<I>,
{
let mut res = TxAndIterator { tx, iter: None };
let tx = unsafe { NonNull::from(&res.tx).as_ref() };
res.iter = Some(iterfun(tx)?);
Ok(Box::new(res))
}
}
impl<'a, I> Drop for TxAndIterator<'a, I>
where
I: Iterator<Item = IteratorItem<'a>> + 'a,
{
fn drop(&mut self) {
drop(self.iter.take());
}
}
impl<'a, I> Iterator for TxAndIterator<'a, I>
where
I: Iterator<Item = IteratorItem<'a>> + 'a,
{
type Item = Result<(Value, Value)>;
fn next(&mut self) -> Option<Self::Item> {
match self.iter.as_mut().unwrap().next() {
None => None,
Some(Err(e)) => Some(Err(e.into())),
Some(Ok((k, v))) => Some(Ok((k.to_vec(), v.to_vec()))),
}
}
}
// ----
#[cfg(target_pointer_width = "64")]
pub fn recommended_map_size() -> usize {
1usize << 40
}
#[cfg(target_pointer_width = "32")]
pub fn recommended_map_size() -> usize {
warn!("LMDB is not recommended on 32-bit systems, database size will be limited");
1usize << 30
}
+266
View File
@@ -0,0 +1,266 @@
use core::ops::Bound;
use std::cell::Cell;
use std::collections::HashMap;
use std::sync::{Arc, RwLock};
use sled::transaction::{
ConflictableTransactionError, TransactionError, Transactional, TransactionalTree,
UnabortableTransactionError,
};
use crate::{
Db, Error, IDb, ITx, ITxFn, Result, TxError, TxFnResult, TxOpError, TxOpResult, TxResult,
TxValueIter, Value, ValueIter,
};
pub use sled;
// -- err
impl From<sled::Error> for Error {
fn from(e: sled::Error) -> Error {
Error(format!("Sled: {}", e).into())
}
}
impl From<sled::Error> for TxOpError {
fn from(e: sled::Error) -> TxOpError {
TxOpError(e.into())
}
}
// -- db
pub struct SledDb {
db: sled::Db,
trees: RwLock<(Vec<sled::Tree>, HashMap<String, usize>)>,
}
impl SledDb {
pub fn init(db: sled::Db) -> Db {
let s = Self {
db,
trees: RwLock::new((Vec::new(), HashMap::new())),
};
Db(Arc::new(s))
}
fn get_tree(&self, i: usize) -> Result<sled::Tree> {
self.trees
.read()
.unwrap()
.0
.get(i)
.cloned()
.ok_or_else(|| Error("invalid tree id".into()))
}
}
impl IDb for SledDb {
fn engine(&self) -> String {
"Sled".into()
}
fn open_tree(&self, name: &str) -> Result<usize> {
let mut trees = self.trees.write().unwrap();
if let Some(i) = trees.1.get(name) {
Ok(*i)
} else {
let tree = self.db.open_tree(name)?;
let i = trees.0.len();
trees.0.push(tree);
trees.1.insert(name.to_string(), i);
Ok(i)
}
}
fn list_trees(&self) -> Result<Vec<String>> {
let mut trees = vec![];
for name in self.db.tree_names() {
let name = std::str::from_utf8(&name)
.map_err(|e| Error(format!("{}", e).into()))?
.to_string();
if name != "__sled__default" {
trees.push(name);
}
}
Ok(trees)
}
// ----
fn get(&self, tree: usize, key: &[u8]) -> Result<Option<Value>> {
let tree = self.get_tree(tree)?;
let val = tree.get(key)?;
Ok(val.map(|x| x.to_vec()))
}
fn len(&self, tree: usize) -> Result<usize> {
let tree = self.get_tree(tree)?;
Ok(tree.len())
}
fn insert(&self, tree: usize, key: &[u8], value: &[u8]) -> Result<Option<Value>> {
let tree = self.get_tree(tree)?;
let old_val = tree.insert(key, value)?;
Ok(old_val.map(|x| x.to_vec()))
}
fn remove(&self, tree: usize, key: &[u8]) -> Result<Option<Value>> {
let tree = self.get_tree(tree)?;
let old_val = tree.remove(key)?;
Ok(old_val.map(|x| x.to_vec()))
}
fn clear(&self, tree: usize) -> Result<()> {
let tree = self.get_tree(tree)?;
tree.clear()?;
Ok(())
}
fn iter(&self, tree: usize) -> Result<ValueIter<'_>> {
let tree = self.get_tree(tree)?;
Ok(Box::new(tree.iter().map(|v| {
v.map(|(x, y)| (x.to_vec(), y.to_vec())).map_err(Into::into)
})))
}
fn iter_rev(&self, tree: usize) -> Result<ValueIter<'_>> {
let tree = self.get_tree(tree)?;
Ok(Box::new(tree.iter().rev().map(|v| {
v.map(|(x, y)| (x.to_vec(), y.to_vec())).map_err(Into::into)
})))
}
fn range<'r>(
&self,
tree: usize,
low: Bound<&'r [u8]>,
high: Bound<&'r [u8]>,
) -> Result<ValueIter<'_>> {
let tree = self.get_tree(tree)?;
Ok(Box::new(tree.range::<&'r [u8], _>((low, high)).map(|v| {
v.map(|(x, y)| (x.to_vec(), y.to_vec())).map_err(Into::into)
})))
}
fn range_rev<'r>(
&self,
tree: usize,
low: Bound<&'r [u8]>,
high: Bound<&'r [u8]>,
) -> Result<ValueIter<'_>> {
let tree = self.get_tree(tree)?;
Ok(Box::new(tree.range::<&'r [u8], _>((low, high)).rev().map(
|v| v.map(|(x, y)| (x.to_vec(), y.to_vec())).map_err(Into::into),
)))
}
// ----
fn transaction(&self, f: &dyn ITxFn) -> TxResult<(), ()> {
let trees = self.trees.read().unwrap();
let res = trees.0.transaction(|txtrees| {
let mut tx = SledTx {
trees: txtrees,
err: Cell::new(None),
};
match f.try_on(&mut tx) {
TxFnResult::Ok => {
assert!(tx.err.into_inner().is_none());
Ok(())
}
TxFnResult::Abort => {
assert!(tx.err.into_inner().is_none());
Err(ConflictableTransactionError::Abort(()))
}
TxFnResult::DbErr => {
let e = tx.err.into_inner().expect("No DB error");
Err(e.into())
}
}
});
match res {
Ok(()) => Ok(()),
Err(TransactionError::Abort(())) => Err(TxError::Abort(())),
Err(TransactionError::Storage(s)) => Err(TxError::Db(s.into())),
}
}
}
// ----
struct SledTx<'a> {
trees: &'a [TransactionalTree],
err: Cell<Option<UnabortableTransactionError>>,
}
impl<'a> SledTx<'a> {
fn get_tree(&self, i: usize) -> TxOpResult<&TransactionalTree> {
self.trees.get(i).ok_or_else(|| {
TxOpError(Error(
"invalid tree id (it might have been openned after the transaction started)".into(),
))
})
}
fn save_error<R>(
&self,
v: std::result::Result<R, UnabortableTransactionError>,
) -> TxOpResult<R> {
match v {
Ok(x) => Ok(x),
Err(e) => {
let txt = format!("{}", e);
self.err.set(Some(e));
Err(TxOpError(Error(txt.into())))
}
}
}
}
impl<'a> ITx for SledTx<'a> {
fn get(&self, tree: usize, key: &[u8]) -> TxOpResult<Option<Value>> {
let tree = self.get_tree(tree)?;
let tmp = self.save_error(tree.get(key))?;
Ok(tmp.map(|x| x.to_vec()))
}
fn len(&self, _tree: usize) -> TxOpResult<usize> {
unimplemented!(".len() in transaction not supported with Sled backend")
}
fn insert(&mut self, tree: usize, key: &[u8], value: &[u8]) -> TxOpResult<Option<Value>> {
let tree = self.get_tree(tree)?;
let old_val = self.save_error(tree.insert(key, value))?;
Ok(old_val.map(|x| x.to_vec()))
}
fn remove(&mut self, tree: usize, key: &[u8]) -> TxOpResult<Option<Value>> {
let tree = self.get_tree(tree)?;
let old_val = self.save_error(tree.remove(key))?;
Ok(old_val.map(|x| x.to_vec()))
}
fn iter(&self, _tree: usize) -> TxOpResult<TxValueIter<'_>> {
unimplemented!("Iterators in transactions not supported with Sled backend");
}
fn iter_rev(&self, _tree: usize) -> TxOpResult<TxValueIter<'_>> {
unimplemented!("Iterators in transactions not supported with Sled backend");
}
fn range<'r>(
&self,
_tree: usize,
_low: Bound<&'r [u8]>,
_high: Bound<&'r [u8]>,
) -> TxOpResult<TxValueIter<'_>> {
unimplemented!("Iterators in transactions not supported with Sled backend");
}
fn range_rev<'r>(
&self,
_tree: usize,
_low: Bound<&'r [u8]>,
_high: Bound<&'r [u8]>,
) -> TxOpResult<TxValueIter<'_>> {
unimplemented!("Iterators in transactions not supported with Sled backend");
}
}
+508
View File
@@ -0,0 +1,508 @@
use core::ops::Bound;
use std::borrow::BorrowMut;
use std::marker::PhantomPinned;
use std::pin::Pin;
use std::ptr::NonNull;
use std::sync::{Arc, Mutex, MutexGuard};
use rusqlite::{params, Connection, Rows, Statement, Transaction};
use crate::{
Db, Error, IDb, ITx, ITxFn, Result, TxError, TxFnResult, TxOpError, TxOpResult, TxResult,
TxValueIter, Value, ValueIter,
};
pub use rusqlite;
// --- err
impl From<rusqlite::Error> for Error {
fn from(e: rusqlite::Error) -> Error {
Error(format!("Sqlite: {}", e).into())
}
}
impl From<rusqlite::Error> for TxOpError {
fn from(e: rusqlite::Error) -> TxOpError {
TxOpError(e.into())
}
}
// -- db
pub struct SqliteDb(Mutex<SqliteDbInner>);
struct SqliteDbInner {
db: Connection,
trees: Vec<String>,
}
impl SqliteDb {
pub fn init(db: rusqlite::Connection) -> Db {
let s = Self(Mutex::new(SqliteDbInner {
db,
trees: Vec::new(),
}));
Db(Arc::new(s))
}
}
impl SqliteDbInner {
fn get_tree(&self, i: usize) -> Result<&'_ str> {
self.trees
.get(i)
.map(String::as_str)
.ok_or_else(|| Error("invalid tree id".into()))
}
fn internal_get(&self, tree: &str, key: &[u8]) -> Result<Option<Value>> {
let mut stmt = self
.db
.prepare(&format!("SELECT v FROM {} WHERE k = ?1", tree))?;
let mut res_iter = stmt.query([key])?;
match res_iter.next()? {
None => Ok(None),
Some(v) => Ok(Some(v.get::<_, Vec<u8>>(0)?)),
}
}
}
impl IDb for SqliteDb {
fn engine(&self) -> String {
format!("sqlite3 v{} (using rusqlite crate)", rusqlite::version())
}
fn open_tree(&self, name: &str) -> Result<usize> {
let name = format!("tree_{}", name.replace(':', "_COLON_"));
let mut this = self.0.lock().unwrap();
if let Some(i) = this.trees.iter().position(|x| x == &name) {
Ok(i)
} else {
trace!("create table {}", name);
this.db.execute(
&format!(
"CREATE TABLE IF NOT EXISTS {} (
k BLOB PRIMARY KEY,
v BLOB
)",
name
),
[],
)?;
trace!("table created: {}, unlocking", name);
let i = this.trees.len();
this.trees.push(name.to_string());
Ok(i)
}
}
fn list_trees(&self) -> Result<Vec<String>> {
let mut trees = vec![];
trace!("list_trees: lock db");
let this = self.0.lock().unwrap();
trace!("list_trees: lock acquired");
let mut stmt = this.db.prepare(
"SELECT name FROM sqlite_schema WHERE type = 'table' AND name LIKE 'tree_%'",
)?;
let mut rows = stmt.query([])?;
while let Some(row) = rows.next()? {
let name = row.get::<_, String>(0)?;
let name = name.replace("_COLON_", ":");
let name = name.strip_prefix("tree_").unwrap().to_string();
trees.push(name);
}
Ok(trees)
}
// ----
fn get(&self, tree: usize, key: &[u8]) -> Result<Option<Value>> {
trace!("get {}: lock db", tree);
let this = self.0.lock().unwrap();
trace!("get {}: lock acquired", tree);
let tree = this.get_tree(tree)?;
this.internal_get(tree, key)
}
fn len(&self, tree: usize) -> Result<usize> {
trace!("len {}: lock db", tree);
let this = self.0.lock().unwrap();
trace!("len {}: lock acquired", tree);
let tree = this.get_tree(tree)?;
let mut stmt = this.db.prepare(&format!("SELECT COUNT(*) FROM {}", tree))?;
let mut res_iter = stmt.query([])?;
match res_iter.next()? {
None => Ok(0),
Some(v) => Ok(v.get::<_, usize>(0)?),
}
}
fn insert(&self, tree: usize, key: &[u8], value: &[u8]) -> Result<Option<Value>> {
trace!("insert {}: lock db", tree);
let this = self.0.lock().unwrap();
trace!("insert {}: lock acquired", tree);
let tree = this.get_tree(tree)?;
let old_val = this.internal_get(tree, key)?;
let sql = match &old_val {
Some(_) => format!("UPDATE {} SET v = ?2 WHERE k = ?1", tree),
None => format!("INSERT INTO {} (k, v) VALUES (?1, ?2)", tree),
};
let n = this.db.execute(&sql, params![key, value])?;
assert_eq!(n, 1);
Ok(old_val)
}
fn remove(&self, tree: usize, key: &[u8]) -> Result<Option<Value>> {
trace!("remove {}: lock db", tree);
let this = self.0.lock().unwrap();
trace!("remove {}: lock acquired", tree);
let tree = this.get_tree(tree)?;
let old_val = this.internal_get(tree, key)?;
if old_val.is_some() {
let n = this
.db
.execute(&format!("DELETE FROM {} WHERE k = ?1", tree), params![key])?;
assert_eq!(n, 1);
}
Ok(old_val)
}
fn clear(&self, tree: usize) -> Result<()> {
trace!("clear {}: lock db", tree);
let this = self.0.lock().unwrap();
trace!("clear {}: lock acquired", tree);
let tree = this.get_tree(tree)?;
this.db.execute(&format!("DELETE FROM {}", tree), [])?;
Ok(())
}
fn iter(&self, tree: usize) -> Result<ValueIter<'_>> {
trace!("iter {}: lock db", tree);
let this = self.0.lock().unwrap();
trace!("iter {}: lock acquired", tree);
let tree = this.get_tree(tree)?;
let sql = format!("SELECT k, v FROM {} ORDER BY k ASC", tree);
DbValueIterator::make(this, &sql, [])
}
fn iter_rev(&self, tree: usize) -> Result<ValueIter<'_>> {
trace!("iter_rev {}: lock db", tree);
let this = self.0.lock().unwrap();
trace!("iter_rev {}: lock acquired", tree);
let tree = this.get_tree(tree)?;
let sql = format!("SELECT k, v FROM {} ORDER BY k DESC", tree);
DbValueIterator::make(this, &sql, [])
}
fn range<'r>(
&self,
tree: usize,
low: Bound<&'r [u8]>,
high: Bound<&'r [u8]>,
) -> Result<ValueIter<'_>> {
trace!("range {}: lock db", tree);
let this = self.0.lock().unwrap();
trace!("range {}: lock acquired", tree);
let tree = this.get_tree(tree)?;
let (bounds_sql, params) = bounds_sql(low, high);
let sql = format!("SELECT k, v FROM {} {} ORDER BY k ASC", tree, bounds_sql);
let params = params
.iter()
.map(|x| x as &dyn rusqlite::ToSql)
.collect::<Vec<_>>();
DbValueIterator::make::<&[&dyn rusqlite::ToSql]>(this, &sql, params.as_ref())
}
fn range_rev<'r>(
&self,
tree: usize,
low: Bound<&'r [u8]>,
high: Bound<&'r [u8]>,
) -> Result<ValueIter<'_>> {
trace!("range_rev {}: lock db", tree);
let this = self.0.lock().unwrap();
trace!("range_rev {}: lock acquired", tree);
let tree = this.get_tree(tree)?;
let (bounds_sql, params) = bounds_sql(low, high);
let sql = format!("SELECT k, v FROM {} {} ORDER BY k DESC", tree, bounds_sql);
let params = params
.iter()
.map(|x| x as &dyn rusqlite::ToSql)
.collect::<Vec<_>>();
DbValueIterator::make::<&[&dyn rusqlite::ToSql]>(this, &sql, params.as_ref())
}
// ----
fn transaction(&self, f: &dyn ITxFn) -> TxResult<(), ()> {
trace!("transaction: lock db");
let mut this = self.0.lock().unwrap();
trace!("transaction: lock acquired");
let this_mut_ref: &mut SqliteDbInner = this.borrow_mut();
let mut tx = SqliteTx {
tx: this_mut_ref
.db
.transaction()
.map_err(Error::from)
.map_err(TxError::Db)?,
trees: &this_mut_ref.trees,
};
let res = match f.try_on(&mut tx) {
TxFnResult::Ok => {
tx.tx.commit().map_err(Error::from).map_err(TxError::Db)?;
Ok(())
}
TxFnResult::Abort => {
tx.tx.rollback().map_err(Error::from).map_err(TxError::Db)?;
Err(TxError::Abort(()))
}
TxFnResult::DbErr => {
tx.tx.rollback().map_err(Error::from).map_err(TxError::Db)?;
Err(TxError::Db(Error(
"(this message will be discarded)".into(),
)))
}
};
trace!("transaction done");
res
}
}
// ----
struct SqliteTx<'a> {
tx: Transaction<'a>,
trees: &'a [String],
}
impl<'a> SqliteTx<'a> {
fn get_tree(&self, i: usize) -> TxOpResult<&'_ str> {
self.trees.get(i).map(String::as_ref).ok_or_else(|| {
TxOpError(Error(
"invalid tree id (it might have been openned after the transaction started)".into(),
))
})
}
fn internal_get(&self, tree: &str, key: &[u8]) -> TxOpResult<Option<Value>> {
let mut stmt = self
.tx
.prepare(&format!("SELECT v FROM {} WHERE k = ?1", tree))?;
let mut res_iter = stmt.query([key])?;
match res_iter.next()? {
None => Ok(None),
Some(v) => Ok(Some(v.get::<_, Vec<u8>>(0)?)),
}
}
}
impl<'a> ITx for SqliteTx<'a> {
fn get(&self, tree: usize, key: &[u8]) -> TxOpResult<Option<Value>> {
let tree = self.get_tree(tree)?;
self.internal_get(tree, key)
}
fn len(&self, tree: usize) -> TxOpResult<usize> {
let tree = self.get_tree(tree)?;
let mut stmt = self.tx.prepare(&format!("SELECT COUNT(*) FROM {}", tree))?;
let mut res_iter = stmt.query([])?;
match res_iter.next()? {
None => Ok(0),
Some(v) => Ok(v.get::<_, usize>(0)?),
}
}
fn insert(&mut self, tree: usize, key: &[u8], value: &[u8]) -> TxOpResult<Option<Value>> {
let tree = self.get_tree(tree)?;
let old_val = self.internal_get(tree, key)?;
let sql = match &old_val {
Some(_) => format!("UPDATE {} SET v = ?2 WHERE k = ?1", tree),
None => format!("INSERT INTO {} (k, v) VALUES (?1, ?2)", tree),
};
let n = self.tx.execute(&sql, params![key, value])?;
assert_eq!(n, 1);
Ok(old_val)
}
fn remove(&mut self, tree: usize, key: &[u8]) -> TxOpResult<Option<Value>> {
let tree = self.get_tree(tree)?;
let old_val = self.internal_get(tree, key)?;
if old_val.is_some() {
let n = self
.tx
.execute(&format!("DELETE FROM {} WHERE k = ?1", tree), params![key])?;
assert_eq!(n, 1);
}
Ok(old_val)
}
fn iter(&self, _tree: usize) -> TxOpResult<TxValueIter<'_>> {
unimplemented!();
}
fn iter_rev(&self, _tree: usize) -> TxOpResult<TxValueIter<'_>> {
unimplemented!();
}
fn range<'r>(
&self,
_tree: usize,
_low: Bound<&'r [u8]>,
_high: Bound<&'r [u8]>,
) -> TxOpResult<TxValueIter<'_>> {
unimplemented!();
}
fn range_rev<'r>(
&self,
_tree: usize,
_low: Bound<&'r [u8]>,
_high: Bound<&'r [u8]>,
) -> TxOpResult<TxValueIter<'_>> {
unimplemented!();
}
}
// ----
struct DbValueIterator<'a> {
db: MutexGuard<'a, SqliteDbInner>,
stmt: Option<Statement<'a>>,
iter: Option<Rows<'a>>,
_pin: PhantomPinned,
}
impl<'a> DbValueIterator<'a> {
fn make<P: rusqlite::Params>(
db: MutexGuard<'a, SqliteDbInner>,
sql: &str,
args: P,
) -> Result<ValueIter<'a>> {
let res = DbValueIterator {
db,
stmt: None,
iter: None,
_pin: PhantomPinned,
};
let mut boxed = Box::pin(res);
trace!("make iterator with sql: {}", sql);
unsafe {
let db = NonNull::from(&boxed.db);
let stmt = db.as_ref().db.prepare(sql)?;
let mut_ref: Pin<&mut DbValueIterator<'a>> = Pin::as_mut(&mut boxed);
Pin::get_unchecked_mut(mut_ref).stmt = Some(stmt);
let mut stmt = NonNull::from(&boxed.stmt);
let iter = stmt.as_mut().as_mut().unwrap().query(args)?;
let mut_ref: Pin<&mut DbValueIterator<'a>> = Pin::as_mut(&mut boxed);
Pin::get_unchecked_mut(mut_ref).iter = Some(iter);
}
Ok(Box::new(DbValueIteratorPin(boxed)))
}
}
impl<'a> Drop for DbValueIterator<'a> {
fn drop(&mut self) {
trace!("drop iter");
drop(self.iter.take());
drop(self.stmt.take());
}
}
struct DbValueIteratorPin<'a>(Pin<Box<DbValueIterator<'a>>>);
impl<'a> Iterator for DbValueIteratorPin<'a> {
type Item = Result<(Value, Value)>;
fn next(&mut self) -> Option<Self::Item> {
let next = unsafe {
let mut_ref: Pin<&mut DbValueIterator<'a>> = Pin::as_mut(&mut self.0);
Pin::get_unchecked_mut(mut_ref).iter.as_mut()?.next()
};
let row = match next {
Err(e) => return Some(Err(e.into())),
Ok(None) => return None,
Ok(Some(r)) => r,
};
let k = match row.get::<_, Vec<u8>>(0) {
Err(e) => return Some(Err(e.into())),
Ok(x) => x,
};
let v = match row.get::<_, Vec<u8>>(1) {
Err(e) => return Some(Err(e.into())),
Ok(y) => y,
};
Some(Ok((k, v)))
}
}
// ----
fn bounds_sql<'r>(low: Bound<&'r [u8]>, high: Bound<&'r [u8]>) -> (String, Vec<Vec<u8>>) {
let mut sql = String::new();
let mut params: Vec<Vec<u8>> = vec![];
match low {
Bound::Included(b) => {
sql.push_str(" WHERE k >= ?1");
params.push(b.to_vec());
}
Bound::Excluded(b) => {
sql.push_str(" WHERE k > ?1");
params.push(b.to_vec());
}
Bound::Unbounded => (),
};
match high {
Bound::Included(b) => {
if !params.is_empty() {
sql.push_str(" AND k <= ?2");
} else {
sql.push_str(" WHERE k <= ?1");
}
params.push(b.to_vec());
}
Bound::Excluded(b) => {
if !params.is_empty() {
sql.push_str(" AND k < ?2");
} else {
sql.push_str(" WHERE k < ?1");
}
params.push(b.to_vec());
}
Bound::Unbounded => (),
}
(sql, params)
}
+106
View File
@@ -0,0 +1,106 @@
use crate::*;
use crate::lmdb_adapter::LmdbDb;
use crate::sled_adapter::SledDb;
use crate::sqlite_adapter::SqliteDb;
fn test_suite(db: Db) {
let tree = db.open_tree("tree").unwrap();
let ka: &[u8] = &b"test"[..];
let kb: &[u8] = &b"zwello"[..];
let kint: &[u8] = &b"tz"[..];
let va: &[u8] = &b"plop"[..];
let vb: &[u8] = &b"plip"[..];
let vc: &[u8] = &b"plup"[..];
assert!(tree.insert(ka, va).unwrap().is_none());
assert_eq!(tree.get(ka).unwrap().unwrap(), va);
let res = db.transaction::<_, (), _>(|mut tx| {
assert_eq!(tx.get(&tree, ka).unwrap().unwrap(), va);
assert_eq!(tx.insert(&tree, ka, vb).unwrap().unwrap(), va);
assert_eq!(tx.get(&tree, ka).unwrap().unwrap(), vb);
tx.commit(12)
});
assert!(matches!(res, Ok(12)));
assert_eq!(tree.get(ka).unwrap().unwrap(), vb);
let res = db.transaction::<(), _, _>(|mut tx| {
assert_eq!(tx.get(&tree, ka).unwrap().unwrap(), vb);
assert_eq!(tx.insert(&tree, ka, vc).unwrap().unwrap(), vb);
assert_eq!(tx.get(&tree, ka).unwrap().unwrap(), vc);
tx.abort(42)
});
assert!(matches!(res, Err(TxError::Abort(42))));
assert_eq!(tree.get(ka).unwrap().unwrap(), vb);
let mut iter = tree.iter().unwrap();
let next = iter.next().unwrap().unwrap();
assert_eq!((next.0.as_ref(), next.1.as_ref()), (ka, vb));
assert!(iter.next().is_none());
drop(iter);
assert!(tree.insert(kb, vc).unwrap().is_none());
assert_eq!(tree.get(kb).unwrap().unwrap(), vc);
let mut iter = tree.iter().unwrap();
let next = iter.next().unwrap().unwrap();
assert_eq!((next.0.as_ref(), next.1.as_ref()), (ka, vb));
let next = iter.next().unwrap().unwrap();
assert_eq!((next.0.as_ref(), next.1.as_ref()), (kb, vc));
assert!(iter.next().is_none());
drop(iter);
let mut iter = tree.range(kint..).unwrap();
let next = iter.next().unwrap().unwrap();
assert_eq!((next.0.as_ref(), next.1.as_ref()), (kb, vc));
assert!(iter.next().is_none());
drop(iter);
let mut iter = tree.range_rev(..kint).unwrap();
let next = iter.next().unwrap().unwrap();
assert_eq!((next.0.as_ref(), next.1.as_ref()), (ka, vb));
assert!(iter.next().is_none());
drop(iter);
let mut iter = tree.iter_rev().unwrap();
let next = iter.next().unwrap().unwrap();
assert_eq!((next.0.as_ref(), next.1.as_ref()), (kb, vc));
let next = iter.next().unwrap().unwrap();
assert_eq!((next.0.as_ref(), next.1.as_ref()), (ka, vb));
assert!(iter.next().is_none());
drop(iter);
}
#[test]
fn test_lmdb_db() {
let path = mktemp::Temp::new_dir().unwrap();
let db = heed::EnvOpenOptions::new()
.max_dbs(100)
.open(&path)
.unwrap();
let db = LmdbDb::init(db);
test_suite(db);
drop(path);
}
#[test]
fn test_sled_db() {
let path = mktemp::Temp::new_dir().unwrap();
let db = SledDb::init(sled::open(path.to_path_buf()).unwrap());
test_suite(db);
drop(path);
}
#[test]
fn test_sqlite_db() {
let db = SqliteDb::init(rusqlite::Connection::open_in_memory().unwrap());
test_suite(db);
}
+4 -2
View File
@@ -21,7 +21,9 @@ path = "tests/lib.rs"
# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
[dependencies]
garage_db = { version = "0.8.0", path = "../db" }
garage_api = { version = "0.7.0", path = "../api" }
garage_block = { version = "0.7.0", path = "../block" }
garage_model = { version = "0.7.0", path = "../model" }
garage_rpc = { version = "0.7.0", path = "../rpc" }
garage_table = { version = "0.7.0", path = "../table" }
@@ -29,6 +31,8 @@ garage_util = { version = "0.7.0", path = "../util" }
garage_web = { version = "0.7.0", path = "../web" }
bytes = "1.0"
bytesize = "1.1"
timeago = "0.3"
hex = "0.4"
tracing = { version = "0.1.30", features = ["log-always"] }
pretty_env_logger = "0.4"
@@ -36,8 +40,6 @@ rand = "0.8"
async-trait = "0.1.7"
sodiumoxide = { version = "0.2.5-0", package = "kuska-sodiumoxide" }
sled = "0.34"
rmp-serde = "0.15"
serde = { version = "1.0", default-features = false, features = ["derive", "rc"] }
serde_bytes = "0.11"
+124 -28
View File
@@ -24,24 +24,34 @@ use garage_model::migrate::Migrate;
use garage_model::permission::*;
use crate::cli::*;
use crate::repair::Repair;
use crate::repair::online::launch_online_repair;
pub const ADMIN_RPC_PATH: &str = "garage/admin_rpc.rs/Rpc";
#[derive(Debug, Serialize, Deserialize)]
#[allow(clippy::large_enum_variant)]
pub enum AdminRpc {
BucketOperation(BucketOperation),
KeyOperation(KeyOperation),
LaunchRepair(RepairOpt),
Migrate(MigrateOpt),
Stats(StatsOpt),
Worker(WorkerOpt),
// Replies
Ok(String),
BucketList(Vec<Bucket>),
BucketInfo(Bucket, HashMap<String, Key>),
BucketInfo {
bucket: Bucket,
relevant_keys: HashMap<String, Key>,
counters: HashMap<String, i64>,
},
KeyList(Vec<(String, String)>),
KeyInfo(Key, HashMap<Uuid, Bucket>),
WorkerList(
HashMap<usize, garage_util::background::WorkerInfo>,
WorkerListOpt,
),
}
impl Rpc for AdminRpc {
@@ -72,6 +82,7 @@ impl AdminRpcHandler {
BucketOperation::Allow(query) => self.handle_bucket_allow(query).await,
BucketOperation::Deny(query) => self.handle_bucket_deny(query).await,
BucketOperation::Website(query) => self.handle_bucket_website(query).await,
BucketOperation::SetQuotas(query) => self.handle_bucket_set_quotas(query).await,
}
}
@@ -87,6 +98,7 @@ impl AdminRpcHandler {
EnumerationOrder::Forward,
)
.await?;
Ok(AdminRpc::BucketList(buckets))
}
@@ -104,6 +116,15 @@ impl AdminRpcHandler {
.get_existing_bucket(bucket_id)
.await?;
let counters = self
.garage
.object_counter_table
.table
.get(&bucket_id, &EmptyKey)
.await?
.map(|x| x.filtered_values(&self.garage.system.ring.borrow()))
.unwrap_or_default();
let mut relevant_keys = HashMap::new();
for (k, _) in bucket
.state
@@ -139,7 +160,11 @@ impl AdminRpcHandler {
}
}
Ok(AdminRpc::BucketInfo(bucket, relevant_keys))
Ok(AdminRpc::BucketInfo {
bucket,
relevant_keys,
counters,
})
}
#[allow(clippy::ptr_arg)]
@@ -431,6 +456,60 @@ impl AdminRpcHandler {
Ok(AdminRpc::Ok(msg))
}
async fn handle_bucket_set_quotas(&self, query: &SetQuotasOpt) -> Result<AdminRpc, Error> {
let bucket_id = self
.garage
.bucket_helper()
.resolve_global_bucket_name(&query.bucket)
.await?
.ok_or_bad_request("Bucket not found")?;
let mut bucket = self
.garage
.bucket_helper()
.get_existing_bucket(bucket_id)
.await?;
let bucket_state = bucket.state.as_option_mut().unwrap();
if query.max_size.is_none() && query.max_objects.is_none() {
return Err(Error::BadRequest(
"You must specify either --max-size or --max-objects (or both) for this command to do something.".to_string(),
));
}
let mut quotas = bucket_state.quotas.get().clone();
match query.max_size.as_ref().map(String::as_ref) {
Some("none") => quotas.max_size = None,
Some(v) => {
let bs = v
.parse::<bytesize::ByteSize>()
.ok_or_bad_request(format!("Invalid size specified: {}", v))?;
quotas.max_size = Some(bs.as_u64());
}
_ => (),
}
match query.max_objects.as_ref().map(String::as_ref) {
Some("none") => quotas.max_objects = None,
Some(v) => {
let mo = v
.parse::<u64>()
.ok_or_bad_request(format!("Invalid number specified: {}", v))?;
quotas.max_objects = Some(mo);
}
_ => (),
}
bucket_state.quotas.update(quotas);
self.garage.bucket_table.insert(&bucket).await?;
Ok(AdminRpc::Ok(format!(
"Quotas updated for {}",
&query.bucket
)))
}
async fn handle_key_cmd(&self, cmd: &KeyOperation) -> Result<AdminRpc, Error> {
match cmd {
KeyOperation::List => self.handle_list_keys().await,
@@ -619,15 +698,7 @@ impl AdminRpcHandler {
)))
}
} else {
let repair = Repair {
garage: self.garage.clone(),
};
self.garage
.system
.background
.spawn_worker("Repair worker".into(), move |must_exit| async move {
repair.repair_worker(opt, must_exit).await
});
launch_online_repair(self.garage.clone(), opt).await;
Ok(AdminRpc::Ok(format!(
"Repair launched on {:?}",
self.garage.system.id
@@ -660,11 +731,11 @@ impl AdminRpcHandler {
}
Ok(AdminRpc::Ok(ret))
} else {
Ok(AdminRpc::Ok(self.gather_stats_local(opt)))
Ok(AdminRpc::Ok(self.gather_stats_local(opt)?))
}
}
fn gather_stats_local(&self, opt: StatsOpt) -> String {
fn gather_stats_local(&self, opt: StatsOpt) -> Result<String, Error> {
let mut ret = String::new();
writeln!(
&mut ret,
@@ -672,6 +743,7 @@ impl AdminRpcHandler {
self.garage.system.garage_version(),
)
.unwrap();
writeln!(&mut ret, "\nDatabase engine: {}", self.garage.db.engine()).unwrap();
// Gather ring statistics
let ring = self.garage.system.ring.borrow().clone();
@@ -689,59 +761,82 @@ impl AdminRpcHandler {
writeln!(&mut ret, " {:?} {}", n, c).unwrap();
}
self.gather_table_stats(&mut ret, &self.garage.bucket_table, &opt);
self.gather_table_stats(&mut ret, &self.garage.key_table, &opt);
self.gather_table_stats(&mut ret, &self.garage.object_table, &opt);
self.gather_table_stats(&mut ret, &self.garage.version_table, &opt);
self.gather_table_stats(&mut ret, &self.garage.block_ref_table, &opt);
self.gather_table_stats(&mut ret, &self.garage.bucket_table, &opt)?;
self.gather_table_stats(&mut ret, &self.garage.key_table, &opt)?;
self.gather_table_stats(&mut ret, &self.garage.object_table, &opt)?;
self.gather_table_stats(&mut ret, &self.garage.version_table, &opt)?;
self.gather_table_stats(&mut ret, &self.garage.block_ref_table, &opt)?;
writeln!(&mut ret, "\nBlock manager stats:").unwrap();
if opt.detailed {
writeln!(
&mut ret,
" number of RC entries (~= number of blocks): {}",
self.garage.block_manager.rc_len()
self.garage.block_manager.rc_len()?
)
.unwrap();
}
writeln!(
&mut ret,
" resync queue length: {}",
self.garage.block_manager.resync_queue_len()
self.garage.block_manager.resync_queue_len()?
)
.unwrap();
writeln!(
&mut ret,
" blocks with resync errors: {}",
self.garage.block_manager.resync_errors_len()
self.garage.block_manager.resync_errors_len()?
)
.unwrap();
ret
Ok(ret)
}
fn gather_table_stats<F, R>(&self, to: &mut String, t: &Arc<Table<F, R>>, opt: &StatsOpt)
fn gather_table_stats<F, R>(
&self,
to: &mut String,
t: &Arc<Table<F, R>>,
opt: &StatsOpt,
) -> Result<(), Error>
where
F: TableSchema + 'static,
R: TableReplication + 'static,
{
writeln!(to, "\nTable stats for {}", F::TABLE_NAME).unwrap();
if opt.detailed {
writeln!(to, " number of items: {}", t.data.store.len()).unwrap();
writeln!(
to,
" number of items: {}",
t.data.store.len().map_err(GarageError::from)?
)
.unwrap();
writeln!(
to,
" Merkle tree size: {}",
t.merkle_updater.merkle_tree_len()
t.merkle_updater.merkle_tree_len()?
)
.unwrap();
}
writeln!(
to,
" Merkle updater todo queue length: {}",
t.merkle_updater.todo_len()
t.merkle_updater.todo_len()?
)
.unwrap();
writeln!(to, " GC todo queue length: {}", t.data.gc_todo_len()).unwrap();
writeln!(to, " GC todo queue length: {}", t.data.gc_todo_len()?).unwrap();
Ok(())
}
// ----
async fn handle_worker_cmd(&self, opt: WorkerOpt) -> Result<AdminRpc, Error> {
match opt.cmd {
WorkerCmd::List { opt } => {
let workers = self.garage.background.get_worker_info();
Ok(AdminRpc::WorkerList(workers, opt))
}
}
}
}
@@ -758,6 +853,7 @@ impl EndpointHandler<AdminRpc> for AdminRpcHandler {
AdminRpc::Migrate(opt) => self.handle_migrate(opt.clone()).await,
AdminRpc::LaunchRepair(opt) => self.handle_launch_repair(opt.clone()).await,
AdminRpc::Stats(opt) => self.handle_stats(opt.clone()).await,
AdminRpc::Worker(opt) => self.handle_worker_cmd(opt.clone()).await,
m => Err(GarageError::unexpected_rpc_message(m).into()),
}
}
+13 -3
View File
@@ -1,4 +1,5 @@
use std::collections::HashSet;
use std::time::Duration;
use garage_util::error::*;
use garage_util::formater::format_table;
@@ -39,6 +40,7 @@ pub async fn cli_command_dispatch(
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,
Command::Worker(wo) => cmd_admin(admin_rpc_endpoint, rpc_host, AdminRpc::Worker(wo)).await,
_ => unreachable!(),
}
}
@@ -100,6 +102,7 @@ pub async fn cmd_status(rpc_cli: &Endpoint<SystemRpc, ()>, rpc_host: NodeID) ->
vec!["ID\tHostname\tAddress\tTags\tZone\tCapacity\tLast seen".to_string()];
for adv in status.iter().filter(|adv| !adv.is_up) {
if let Some(NodeRoleV(Some(cfg))) = layout.roles.get(&adv.id) {
let tf = timeago::Formatter::new();
failed_nodes.push(format!(
"{id:?}\t{host}\t{addr}\t[{tags}]\t{zone}\t{capacity}\t{last_seen}",
id = adv.id,
@@ -110,7 +113,7 @@ pub async fn cmd_status(rpc_cli: &Endpoint<SystemRpc, ()>, rpc_host: NodeID) ->
capacity = cfg.capacity_string(),
last_seen = adv
.last_seen_secs_ago
.map(|s| format!("{}s ago", s))
.map(|s| tf.convert(Duration::from_secs(s)))
.unwrap_or_else(|| "never seen".into()),
));
}
@@ -169,8 +172,12 @@ pub async fn cmd_admin(
AdminRpc::BucketList(bl) => {
print_bucket_list(bl);
}
AdminRpc::BucketInfo(bucket, rk) => {
print_bucket_info(&bucket, &rk);
AdminRpc::BucketInfo {
bucket,
relevant_keys,
counters,
} => {
print_bucket_info(&bucket, &relevant_keys, &counters);
}
AdminRpc::KeyList(kl) => {
print_key_list(kl);
@@ -178,6 +185,9 @@ pub async fn cmd_admin(
AdminRpc::KeyInfo(key, rb) => {
print_key_info(&key, &rb);
}
AdminRpc::WorkerList(wi, wlo) => {
print_worker_info(wi, wlo);
}
r => {
error!("Unexpected response: {:?}", r);
}
+99 -3
View File
@@ -33,13 +33,22 @@ pub enum Command {
#[structopt(name = "migrate")]
Migrate(MigrateOpt),
/// Start repair of node data
/// Start repair of node data on remote node
#[structopt(name = "repair")]
Repair(RepairOpt),
/// Offline reparation of node data (these repairs must be run offline
/// directly on the server node)
#[structopt(name = "offline-repair")]
OfflineRepair(OfflineRepairOpt),
/// Gather node statistics
#[structopt(name = "stats")]
Stats(StatsOpt),
/// Manage background workers
#[structopt(name = "worker")]
Worker(WorkerOpt),
}
#[derive(StructOpt, Debug)]
@@ -175,6 +184,10 @@ pub enum BucketOperation {
/// Expose as website or not
#[structopt(name = "website")]
Website(WebsiteOpt),
/// Set the quotas for this bucket
#[structopt(name = "set-quotas")]
SetQuotas(SetQuotasOpt),
}
#[derive(Serialize, Deserialize, StructOpt, Debug)]
@@ -261,6 +274,21 @@ pub struct PermBucketOpt {
pub bucket: String,
}
#[derive(Serialize, Deserialize, StructOpt, Debug)]
pub struct SetQuotasOpt {
/// Bucket name
pub bucket: String,
/// Set a maximum size for the bucket (specify a size e.g. in MiB or GiB,
/// or `none` for no size restriction)
#[structopt(long = "max-size")]
pub max_size: Option<String>,
/// Set a maximum number of objects for the bucket (or `none` for no restriction)
#[structopt(long = "max-objects")]
pub max_objects: Option<String>,
}
#[derive(Serialize, Deserialize, StructOpt, Debug)]
pub enum KeyOperation {
/// List keys
@@ -399,12 +427,54 @@ pub enum RepairWhat {
/// 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")]
#[structopt(subcommand)]
cmd: ScrubCmd,
},
}
#[derive(Serialize, Deserialize, StructOpt, Debug, Eq, PartialEq, Clone)]
pub enum ScrubCmd {
/// Start scrub
#[structopt(name = "start")]
Start,
/// Pause scrub (it will resume automatically after 24 hours)
#[structopt(name = "pause")]
Pause,
/// Resume paused scrub
#[structopt(name = "resume")]
Resume,
/// Cancel scrub in progress
#[structopt(name = "cancel")]
Cancel,
/// Set tranquility level for in-progress and future scrubs
#[structopt(name = "set-tranquility")]
SetTranquility {
#[structopt()]
tranquility: u32,
},
}
#[derive(Serialize, Deserialize, StructOpt, Debug, Clone)]
pub struct OfflineRepairOpt {
/// Confirm the launch of the repair operation
#[structopt(long = "yes")]
pub yes: bool,
#[structopt(subcommand)]
pub what: OfflineRepairWhat,
}
#[derive(Serialize, Deserialize, StructOpt, Debug, Eq, PartialEq, Clone)]
pub enum OfflineRepairWhat {
/// Repair K2V item counters
#[cfg(feature = "k2v")]
#[structopt(name = "k2v_item_counters")]
K2VItemCounters,
/// Repair object counters
#[structopt(name = "object_counters")]
ObjectCounters,
}
#[derive(Serialize, Deserialize, StructOpt, Debug, Clone)]
pub struct StatsOpt {
/// Gather statistics from all nodes
@@ -415,3 +485,29 @@ pub struct StatsOpt {
#[structopt(short = "d", long = "detailed")]
pub detailed: bool,
}
#[derive(Serialize, Deserialize, StructOpt, Debug, Clone)]
pub struct WorkerOpt {
#[structopt(subcommand)]
pub cmd: WorkerCmd,
}
#[derive(Serialize, Deserialize, StructOpt, Debug, Eq, PartialEq, Clone)]
pub enum WorkerCmd {
/// List all workers on Garage node
#[structopt(name = "list")]
List {
#[structopt(flatten)]
opt: WorkerListOpt,
},
}
#[derive(Serialize, Deserialize, StructOpt, Debug, Eq, PartialEq, Clone, Copy)]
pub struct WorkerListOpt {
/// Show only busy workers
#[structopt(short = "b", long = "busy")]
pub busy: bool,
/// Show only workers with errors
#[structopt(short = "e", long = "errors")]
pub errors: bool,
}
+102 -3
View File
@@ -1,12 +1,18 @@
use std::collections::HashMap;
use std::time::Duration;
use garage_util::background::*;
use garage_util::crdt::*;
use garage_util::data::Uuid;
use garage_util::error::*;
use garage_util::formater::format_table;
use garage_util::time::*;
use garage_model::bucket_table::*;
use garage_model::key_table::*;
use garage_model::s3::object_table::{BYTES, OBJECTS, UNFINISHED_UPLOADS};
use crate::cli::structs::WorkerListOpt;
pub fn print_bucket_list(bl: Vec<Bucket>) {
println!("List of buckets:");
@@ -29,11 +35,12 @@ pub fn print_bucket_list(bl: Vec<Bucket>) {
[((k, n), _, _)] => format!("{}:{}", k, n),
s => format!("[{} local aliases]", s.len()),
};
table.push(format!(
"\t{}\t{}\t{}",
aliases.join(","),
local_aliases_n,
hex::encode(bucket.id)
hex::encode(bucket.id),
));
}
format_table(table);
@@ -121,7 +128,11 @@ pub fn print_key_info(key: &Key, relevant_buckets: &HashMap<Uuid, Bucket>) {
}
}
pub fn print_bucket_info(bucket: &Bucket, relevant_keys: &HashMap<String, Key>) {
pub fn print_bucket_info(
bucket: &Bucket,
relevant_keys: &HashMap<String, Key>,
counters: &HashMap<String, i64>,
) {
let key_name = |k| {
relevant_keys
.get(k)
@@ -133,7 +144,42 @@ pub fn print_bucket_info(bucket: &Bucket, relevant_keys: &HashMap<String, Key>)
match &bucket.state {
Deletable::Deleted => println!("Bucket is deleted."),
Deletable::Present(p) => {
println!("Website access: {}", p.website_config.get().is_some());
let size =
bytesize::ByteSize::b(counters.get(BYTES).cloned().unwrap_or_default() as u64);
println!(
"\nSize: {} ({})",
size.to_string_as(true),
size.to_string_as(false)
);
println!(
"Objects: {}",
counters.get(OBJECTS).cloned().unwrap_or_default()
);
println!(
"Unfinished multipart uploads: {}",
counters
.get(UNFINISHED_UPLOADS)
.cloned()
.unwrap_or_default()
);
println!("\nWebsite access: {}", p.website_config.get().is_some());
let quotas = p.quotas.get();
if quotas.max_size.is_some() || quotas.max_objects.is_some() {
println!("\nQuotas:");
if let Some(ms) = quotas.max_size {
let ms = bytesize::ByteSize::b(ms);
println!(
" maximum size: {} ({})",
ms.to_string_as(true),
ms.to_string_as(false)
);
}
if let Some(mo) = quotas.max_objects {
println!(" maximum number of objects: {}", mo);
}
}
println!("\nGlobal aliases:");
for (alias, _, active) in p.aliases.items().iter() {
@@ -194,3 +240,56 @@ pub fn find_matching_node(
Ok(candidates[0])
}
}
pub fn print_worker_info(wi: HashMap<usize, WorkerInfo>, wlo: WorkerListOpt) {
let mut wi = wi.into_iter().collect::<Vec<_>>();
wi.sort_by_key(|(tid, info)| {
(
match info.state {
WorkerState::Busy | WorkerState::Throttled(_) => 0,
WorkerState::Idle => 1,
WorkerState::Done => 2,
},
*tid,
)
});
let mut table = vec![];
for (tid, info) in wi.iter() {
if wlo.busy && !matches!(info.state, WorkerState::Busy | WorkerState::Throttled(_)) {
continue;
}
if wlo.errors && info.errors == 0 {
continue;
}
table.push(format!("{}\t{}\t{}", tid, info.state, info.name));
if let Some(i) = &info.info {
table.push(format!("\t\t {}", i));
}
let tf = timeago::Formatter::new();
let (err_ago, err_msg) = info
.last_error
.as_ref()
.map(|(m, t)| {
(
tf.convert(Duration::from_millis(now_msec() - t)),
m.as_str(),
)
})
.unwrap_or(("(?) ago".into(), "(?)"));
if info.consecutive_errors > 0 {
table.push(format!(
"\t\t {} consecutive errors ({} total), last {}",
info.consecutive_errors, info.errors, err_ago,
));
table.push(format!("\t\t {}", err_msg));
} else if info.errors > 0 {
table.push(format!("\t\t ({} errors, last {})", info.errors, err_ago,));
if wlo.errors {
table.push(format!("\t\t {}", err_msg));
}
}
}
format_table(table);
}
+9 -9
View File
@@ -61,17 +61,17 @@ async fn main() {
pretty_env_logger::init();
sodiumoxide::init().expect("Unable to init sodiumoxide");
// 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 opt = Opt::from_args();
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(opt.config_file).await
Command::Server => server::run_server(opt.config_file).await,
Command::OfflineRepair(repair_opt) => {
repair::offline::offline_repair(opt.config_file, repair_opt).await
}
Command::Node(NodeOperation::NodeId(node_id_opt)) => {
node_id_command(opt.config_file, node_id_opt.quiet)
-149
View File
@@ -1,149 +0,0 @@
use std::sync::Arc;
use tokio::sync::watch;
use garage_model::garage::Garage;
use garage_model::s3::block_ref_table::*;
use garage_model::s3::object_table::*;
use garage_model::s3::version_table::*;
use garage_table::*;
use garage_util::error::Error;
use crate::*;
pub struct Repair {
pub garage: Arc<Garage>,
}
impl Repair {
pub async fn repair_worker(&self, opt: RepairOpt, must_exit: watch::Receiver<bool>) {
if let Err(e) = self.repair_worker_aux(opt, must_exit).await {
warn!("Repair worker failed with error: {}", e);
}
}
async fn repair_worker_aux(
&self,
opt: RepairOpt,
must_exit: watch::Receiver<bool>,
) -> Result<(), Error> {
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?;
}
}
Ok(())
}
async fn repair_versions(&self, must_exit: &watch::Receiver<bool>) -> Result<(), Error> {
let mut pos = vec![];
while let Some((item_key, item_bytes)) =
self.garage.version_table.data.store.get_gt(&pos)?
{
pos = item_key.to_vec();
let version = rmp_serde::decode::from_read_ref::<_, Version>(item_bytes.as_ref())?;
if version.deleted.get() {
continue;
}
let object = self
.garage
.object_table
.get(&version.bucket_id, &version.key)
.await?;
let version_exists = match object {
Some(o) => o
.versions()
.iter()
.any(|x| x.uuid == version.uuid && x.state != ObjectVersionState::Aborted),
None => false,
};
if !version_exists {
info!("Repair versions: marking version as deleted: {:?}", version);
self.garage
.version_table
.insert(&Version::new(
version.uuid,
version.bucket_id,
version.key,
true,
))
.await?;
}
if *must_exit.borrow() {
break;
}
}
Ok(())
}
async fn repair_block_ref(&self, must_exit: &watch::Receiver<bool>) -> Result<(), Error> {
let mut pos = vec![];
while let Some((item_key, item_bytes)) =
self.garage.block_ref_table.data.store.get_gt(&pos)?
{
pos = item_key.to_vec();
let block_ref = rmp_serde::decode::from_read_ref::<_, BlockRef>(item_bytes.as_ref())?;
if block_ref.deleted.get() {
continue;
}
let version = self
.garage
.version_table
.get(&block_ref.version, &EmptyKey)
.await?;
// The version might not exist if it has been GC'ed
let ref_exists = version.map(|v| !v.deleted.get()).unwrap_or(false);
if !ref_exists {
info!(
"Repair block ref: marking block_ref as deleted: {:?}",
block_ref
);
self.garage
.block_ref_table
.insert(&BlockRef {
block: block_ref.block,
version: block_ref.version,
deleted: true.into(),
})
.await?;
}
if *must_exit.borrow() {
break;
}
}
Ok(())
}
}
+2
View File
@@ -0,0 +1,2 @@
pub mod offline;
pub mod online;
+55
View File
@@ -0,0 +1,55 @@
use std::path::PathBuf;
use tokio::sync::watch;
use garage_util::background::*;
use garage_util::config::*;
use garage_util::error::*;
use garage_model::garage::Garage;
use crate::cli::structs::*;
pub async fn offline_repair(config_file: PathBuf, opt: OfflineRepairOpt) -> Result<(), Error> {
if !opt.yes {
return Err(Error::Message(
"Please add the --yes flag to launch repair operation".into(),
));
}
info!("Loading configuration...");
let config = read_config(config_file)?;
info!("Initializing background runner...");
let (done_tx, done_rx) = watch::channel(false);
let (background, await_background_done) = BackgroundRunner::new(16, done_rx);
info!("Initializing Garage main data store...");
let garage = Garage::new(config.clone(), background)?;
info!("Launching repair operation...");
match opt.what {
#[cfg(feature = "k2v")]
OfflineRepairWhat::K2VItemCounters => {
garage
.k2v
.counter_table
.offline_recount_all(&garage.k2v.item_table)?;
}
OfflineRepairWhat::ObjectCounters => {
garage
.object_counter_table
.offline_recount_all(&garage.object_table)?;
}
}
info!("Repair operation finished, shutting down Garage internals...");
done_tx.send(true).unwrap();
drop(garage);
await_background_done.await?;
info!("Cleaning up...");
Ok(())
}
+215
View File
@@ -0,0 +1,215 @@
use std::sync::Arc;
use std::time::Duration;
use async_trait::async_trait;
use tokio::sync::watch;
use garage_block::repair::ScrubWorkerCommand;
use garage_model::garage::Garage;
use garage_model::s3::block_ref_table::*;
use garage_model::s3::object_table::*;
use garage_model::s3::version_table::*;
use garage_table::*;
use garage_util::background::*;
use garage_util::error::Error;
use crate::*;
pub async fn launch_online_repair(garage: Arc<Garage>, opt: RepairOpt) {
match opt.what {
RepairWhat::Tables => {
info!("Launching a full sync of tables");
garage.bucket_table.syncer.add_full_sync();
garage.object_table.syncer.add_full_sync();
garage.version_table.syncer.add_full_sync();
garage.block_ref_table.syncer.add_full_sync();
garage.key_table.syncer.add_full_sync();
}
RepairWhat::Versions => {
info!("Repairing the versions table");
garage
.background
.spawn_worker(RepairVersionsWorker::new(garage.clone()));
}
RepairWhat::BlockRefs => {
info!("Repairing the block refs table");
garage
.background
.spawn_worker(RepairBlockrefsWorker::new(garage.clone()));
}
RepairWhat::Blocks => {
info!("Repairing the stored blocks");
garage
.background
.spawn_worker(garage_block::repair::RepairWorker::new(
garage.block_manager.clone(),
));
}
RepairWhat::Scrub { cmd } => {
let cmd = match cmd {
ScrubCmd::Start => ScrubWorkerCommand::Start,
ScrubCmd::Pause => ScrubWorkerCommand::Pause(Duration::from_secs(3600 * 24)),
ScrubCmd::Resume => ScrubWorkerCommand::Resume,
ScrubCmd::Cancel => ScrubWorkerCommand::Cancel,
ScrubCmd::SetTranquility { tranquility } => {
ScrubWorkerCommand::SetTranquility(tranquility)
}
};
info!("Sending command to scrub worker: {:?}", cmd);
garage.block_manager.send_scrub_command(cmd).await;
}
}
}
// ----
struct RepairVersionsWorker {
garage: Arc<Garage>,
pos: Vec<u8>,
counter: usize,
}
impl RepairVersionsWorker {
fn new(garage: Arc<Garage>) -> Self {
Self {
garage,
pos: vec![],
counter: 0,
}
}
}
#[async_trait]
impl Worker for RepairVersionsWorker {
fn name(&self) -> String {
"Version repair worker".into()
}
fn info(&self) -> Option<String> {
Some(format!("{} items done", self.counter))
}
async fn work(&mut self, _must_exit: &mut watch::Receiver<bool>) -> Result<WorkerState, Error> {
let item_bytes = match self.garage.version_table.data.store.get_gt(&self.pos)? {
Some((k, v)) => {
self.pos = k;
v
}
None => {
info!("repair_versions: finished, done {}", self.counter);
return Ok(WorkerState::Done);
}
};
self.counter += 1;
let version = rmp_serde::decode::from_read_ref::<_, Version>(&item_bytes)?;
if !version.deleted.get() {
let object = self
.garage
.object_table
.get(&version.bucket_id, &version.key)
.await?;
let version_exists = match object {
Some(o) => o
.versions()
.iter()
.any(|x| x.uuid == version.uuid && x.state != ObjectVersionState::Aborted),
None => false,
};
if !version_exists {
info!("Repair versions: marking version as deleted: {:?}", version);
self.garage
.version_table
.insert(&Version::new(
version.uuid,
version.bucket_id,
version.key,
true,
))
.await?;
}
}
Ok(WorkerState::Busy)
}
async fn wait_for_work(&mut self, _must_exit: &watch::Receiver<bool>) -> WorkerState {
unreachable!()
}
}
// ----
struct RepairBlockrefsWorker {
garage: Arc<Garage>,
pos: Vec<u8>,
counter: usize,
}
impl RepairBlockrefsWorker {
fn new(garage: Arc<Garage>) -> Self {
Self {
garage,
pos: vec![],
counter: 0,
}
}
}
#[async_trait]
impl Worker for RepairBlockrefsWorker {
fn name(&self) -> String {
"Block refs repair worker".into()
}
fn info(&self) -> Option<String> {
Some(format!("{} items done", self.counter))
}
async fn work(&mut self, _must_exit: &mut watch::Receiver<bool>) -> Result<WorkerState, Error> {
let item_bytes = match self.garage.block_ref_table.data.store.get_gt(&self.pos)? {
Some((k, v)) => {
self.pos = k;
v
}
None => {
info!("repair_block_ref: finished, done {}", self.counter);
return Ok(WorkerState::Done);
}
};
self.counter += 1;
let block_ref = rmp_serde::decode::from_read_ref::<_, BlockRef>(&item_bytes)?;
if !block_ref.deleted.get() {
let version = self
.garage
.version_table
.get(&block_ref.version, &EmptyKey)
.await?;
// The version might not exist if it has been GC'ed
let ref_exists = version.map(|v| !v.deleted.get()).unwrap_or(false);
if !ref_exists {
info!(
"Repair block ref: marking block_ref as deleted: {:?}",
block_ref
);
self.garage
.block_ref_table
.insert(&BlockRef {
block: block_ref.block,
version: block_ref.version,
deleted: true.into(),
})
.await?;
}
}
Ok(WorkerState::Busy)
}
async fn wait_for_work(&mut self, _must_exit: &watch::Receiver<bool>) -> WorkerState {
unreachable!()
}
}
+3 -12
View File
@@ -27,24 +27,14 @@ async fn wait_from(mut chan: watch::Receiver<bool>) {
pub async fn run_server(config_file: PathBuf) -> Result<(), Error> {
info!("Loading configuration...");
let config = read_config(config_file).expect("Unable to read config file");
info!("Opening database...");
let mut db_path = config.metadata_dir.clone();
db_path.push("db");
let db = sled::Config::default()
.path(&db_path)
.cache_capacity(config.sled_cache_capacity)
.flush_every_ms(Some(config.sled_flush_every_ms))
.open()
.expect("Unable to open sled DB");
let config = read_config(config_file)?;
info!("Initializing background runner...");
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);
let garage = Garage::new(config.clone(), background)?;
info!("Initialize tracing...");
if let Some(export_to) = config.admin.trace_sink {
@@ -54,6 +44,7 @@ pub async fn run_server(config_file: PathBuf) -> Result<(), Error> {
info!("Initialize Admin API server and metrics collector...");
let admin_server = AdminApiServer::new(garage.clone());
info!("Launching internal Garage cluster communications...");
let run_system = tokio::spawn(garage.system.clone().run(watch_cancel.clone()));
info!("Create admin RPC handler...");
+3 -5
View File
@@ -29,8 +29,7 @@ async fn test_bucket_all() {
.unwrap()
.iter()
.filter(|x| x.name.as_ref().is_some())
.find(|x| x.name.as_ref().unwrap() == "hello")
.is_some());
.any(|x| x.name.as_ref().unwrap() == "hello"));
}
{
// Get its location
@@ -75,13 +74,12 @@ async fn test_bucket_all() {
{
// Check bucket is deleted with List buckets
let r = ctx.client.list_buckets().send().await.unwrap();
assert!(r
assert!(!r
.buckets
.as_ref()
.unwrap()
.iter()
.filter(|x| x.name.as_ref().is_some())
.find(|x| x.name.as_ref().unwrap() == "hello")
.is_none());
.any(|x| x.name.as_ref().unwrap() == "hello"));
}
}

Some files were not shown because too many files have changed in this diff Show More