Compare commits

..

6 Commits

Author SHA1 Message Date
houseme 6264314cdc Merge branch 'main' into fix/b5-t4-b2-remote-target-serde 2026-08-23 12:28:53 +08:00
Zhengchao An b6ba89d9e4 docs(testing): document CI gate matrix (#6412) 2026-08-23 12:09:06 +08:00
houseme 648d5166e2 feat(allocator): replace mimalloc/libmimalloc-sys with rustfs-mimalloc/rustfs-mimalloc-sys (#6404)
Replace the upstream xonatius/mimalloc_rust.git fork (mimalloc + libmimalloc-sys)
with the published rustfs-mimalloc (v0.5.0) and rustfs-mimalloc-sys (v0.5.0) crates
from crates.io.

The new crates are based on mimalloc V3 (v3.5.0) and provide:
- MiMalloc global allocator with safe API (collect, stats_json, process_info)
- Heap management and arena operations (heap module)
- Full FFI bindings to mimalloc V3

Changes:
- Workspace deps: mimalloc + libmimalloc-sys (git) → rustfs-mimalloc + rustfs-mimalloc-sys (crates.io)
- allocator_reclaim.rs: libmimalloc_sys::mi_collect → rustfs_mimalloc::MiMalloc::collect
- memory_observability.rs: raw FFI mi_stats_get_json → MiMalloc::stats_json()
- main.rs: heap ownership tests use Heap::contains() (V3 API)
- deny.toml: remove xonatius/mimalloc_rust.git from allow-git

Co-authored-by: heihutu <heihutu@gmail.com>
2026-08-23 12:07:25 +08:00
houseme 84eb5aebef fix(ecstore): remove inline write debug noise (#6408)
* fix(ecstore): remove inline write debug noise

Co-Authored-By: heihutu <heihutu@gmail.com>

* fix(ecstore): satisfy warning-as-error lints

Co-Authored-By: heihutu <heihutu@gmail.com>

---------

Co-authored-by: heihutu <heihutu@gmail.com>
2026-08-23 12:07:20 +08:00
唐小鸭 cddc003c16 fix(admin): emit madmin bandwidthlimit tag in list-remote-targets
madmin-go v3.0.109 `BucketTarget` (bucket-targets.go) tags the limit
`json:"bandwidthlimit,omitempty"`, not `bandwidth`, so the previous rename
still left mc decoding a zero bandwidth limit for every nonzero target.
Emit the exact tag in remote_target_admin_json, and add the same
`bandwidthlimit` alias on the ecstore `BucketTarget` reader so a
MinIO-written bucket-targets.json keeps its limit on migration (the legacy
`bandwidth` alias stays for compatibility). The persisted snake_case wire
key and the inbound RemoteTargetRequest aliases are unchanged.

Cover it with a madmin-shaped decode regression: the admin response is
decoded through a mirror of the v3.0.109 json tags and must yield the
nonzero bandwidth limit plus the other renamed fields.
2026-08-23 00:59:45 +08:00
唐小鸭 fbecd38e2d fix(admin): use madmin key names in list-remote-targets response
The admin list-remote-targets response serialized five fields under the
persisted snake_case keys (bandwidth_limit, storage_class, reset_id,
deployment_id, credentials.session_token), which madmin/mc decode to zero
values, blanking the bandwidth and reset-id columns of mc replicate ls.
Rename them to the madmin JSON tags (bandwidth, storageclass, resetID,
deploymentID, sessionToken) in remote_target_admin_json only, following
the existing Go-duration overwrite precedent there.

Symmetrically, BucketTarget deserialization only aliased bandwidth, so a
MinIO-written bucket-targets.json silently dropped the other four fields
to defaults on migration; add serde aliases for them. Aliases and the
response-layer rename leave the persisted bucket-targets.json and the
msgpack struct-map wire format untouched, pinned by new tests.

Refs rustfs/backlog#1951, spec rustfs/backlog#1946.
2026-08-21 18:53:48 +08:00
19 changed files with 484 additions and 222 deletions
+2 -1
View File
@@ -30,7 +30,8 @@ make build-docker BUILD_OS=ubuntu22.04
- Crate membership: `Cargo.toml` `[workspace].members`
- Architecture, layering, crate map: [ARCHITECTURE.md](ARCHITECTURE.md)
- Migration guardrails & readiness contracts: [docs/architecture/](docs/architecture/README.md)
- CI gates: `.github/workflows/ci.yml` (source of truth; never copy its steps into docs)
- CI workflow steps: `.github/workflows/`; event, timeout, and required-status
matrix: [docs/testing/ci-gates.md](docs/testing/ci-gates.md)
- Test-layer taxonomy, per-layer entry commands, serial/nextest rules, flake
policy: [docs/testing/README.md](docs/testing/README.md)
- Tier/ILM transition debugging (xl.meta inspection, versionId tracing):
+2
View File
@@ -70,6 +70,8 @@ make pre-pr
> For the full test-layer taxonomy (unit / ecstore black-box / e2e / s3s-e2e / S3 compatibility / chaos / fuzz / bench), each layer's entry command, the naming conventions the migration gate depends on, and the serial/nextest rules, see [docs/testing/README.md](docs/testing/README.md).
> For the event, timeout, required-status, and local reproduction matrix, see [docs/testing/ci-gates.md](docs/testing/ci-gates.md).
### 🔒 Automated Pre-commit Hooks
#### What `make pre-commit` and `make pre-pr` actually run
Generated
+22 -27
View File
@@ -1858,9 +1858,9 @@ dependencies = [
[[package]]
name = "cc"
version = "1.4.3"
version = "1.4.4"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "509591b7bcd67f4ef775afad7662703b4935daaa6ec0e5605cfb1090b32a2b6d"
checksum = "0ad534f4357a5264cce5019c989cf66a4f0dc4e0d1b1d15f8aacec0ff7360273"
dependencies = [
"find-msvc-tools",
"jobserver",
@@ -2522,12 +2522,6 @@ dependencies = [
"subtle",
]
[[package]]
name = "cty"
version = "0.2.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "b365fabc795046672053e29c954733ec3b05e4be654ab130fe8f1f94d7051f35"
[[package]]
name = "curve25519-dalek"
version = "4.1.3"
@@ -5988,15 +5982,6 @@ version = "0.2.16"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "b6d2cec3eae94f9f509c767b45932f1ada8350c4bdb85af2fcab4a3c14807981"
[[package]]
name = "libmimalloc-sys"
version = "0.1.49"
source = "git+https://github.com/xonatius/mimalloc_rust.git?rev=6d4c41bb10c6d9da1d1b6f07b38c4cc051667f11#6d4c41bb10c6d9da1d1b6f07b38c4cc051667f11"
dependencies = [
"cc",
"cty",
]
[[package]]
name = "libredox"
version = "0.1.20"
@@ -6397,14 +6382,6 @@ dependencies = [
"synstructure 0.13.2",
]
[[package]]
name = "mimalloc"
version = "0.1.52"
source = "git+https://github.com/xonatius/mimalloc_rust.git?rev=6d4c41bb10c6d9da1d1b6f07b38c4cc051667f11#6d4c41bb10c6d9da1d1b6f07b38c4cc051667f11"
dependencies = [
"libmimalloc-sys",
]
[[package]]
name = "mime"
version = "0.3.17"
@@ -9162,13 +9139,11 @@ dependencies = [
"insta",
"jiff",
"libc",
"libmimalloc-sys",
"libsystemd",
"matchit 0.9.2",
"md-5 0.11.0",
"metrics",
"metrics-util",
"mimalloc",
"mime_guess",
"opentelemetry",
"opentelemetry_sdk",
@@ -9204,6 +9179,8 @@ dependencies = [
"rustfs-lock",
"rustfs-log-analyzer",
"rustfs-madmin",
"rustfs-mimalloc",
"rustfs-mimalloc-sys",
"rustfs-notify",
"rustfs-object-capacity",
"rustfs-object-data-cache",
@@ -9875,6 +9852,24 @@ dependencies = [
"tokio",
]
[[package]]
name = "rustfs-mimalloc"
version = "0.5.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "a406f4aa07084301d485beec873af6dccc8e3f8762da244743df92038b1db1a6"
dependencies = [
"rustfs-mimalloc-sys",
]
[[package]]
name = "rustfs-mimalloc-sys"
version = "0.5.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "c3051b819175f58445d4c369a72f0ab88149f3885ba8bea2aff3be01f53fe7cd"
dependencies = [
"cc",
]
[[package]]
name = "rustfs-notify"
version = "1.0.0-rc.3"
+2 -2
View File
@@ -350,8 +350,8 @@ russh-sftp = "2.4.0"
dav-server = "0.11.0"
# Performance Analysis and Memory Profiling
mimalloc = { version = "0.1.52", git = "https://github.com/xonatius/mimalloc_rust.git", rev = "6d4c41bb10c6d9da1d1b6f07b38c4cc051667f11" }
libmimalloc-sys = { version = "0.1.49", git = "https://github.com/xonatius/mimalloc_rust.git", rev = "6d4c41bb10c6d9da1d1b6f07b38c4cc051667f11", features = ["extended"] }
rustfs-mimalloc = { version = "0.5.0" }
rustfs-mimalloc-sys = { version = "0.5.0" }
hotpath = { version = "0.23.3", default-features = false }
# Snapshot testing for output format regression detection
insta = { version = "1.48" }
@@ -32,6 +32,10 @@ pub struct Credentials {
pub access_key: String,
#[serde(rename = "secretKey")]
pub secret_key: String,
// The aliases accept madmin's JSON tags (MinIO-written bucket-targets
// metadata and mc request bodies) without changing the snake_case
// persisted/peer wire format this struct serializes to.
#[serde(alias = "sessionToken")]
pub session_token: Option<String>,
pub expiration: Option<Timestamp>,
}
@@ -202,12 +206,14 @@ pub struct BucketTarget {
#[serde(default)]
pub region: String,
#[serde(alias = "bandwidth", default)]
// madmin-go v3.0.109 tags this `bandwidthlimit`; `bandwidth` is a legacy
// alias kept for inputs written before the madmin tag was verified.
#[serde(alias = "bandwidthlimit", alias = "bandwidth", default)]
pub bandwidth_limit: i64,
#[serde(rename = "replicationSync", default)]
pub replication_sync: bool,
#[serde(default)]
#[serde(alias = "storageclass", default)]
pub storage_class: String,
#[serde(rename = "skipTlsVerify", default)]
pub skip_tls_verify: bool,
@@ -220,7 +226,7 @@ pub struct BucketTarget {
#[serde(rename = "resetBeforeDate", with = "time::serde::rfc3339::option", default)]
pub reset_before_date: Option<OffsetDateTime>,
#[serde(default)]
#[serde(alias = "resetID", default)]
pub reset_id: String,
#[serde(rename = "totalDowntime", with = "duration_seconds", default)]
pub total_downtime: Duration,
@@ -233,7 +239,7 @@ pub struct BucketTarget {
#[serde(default)]
pub latency: LatencyStat,
#[serde(default)]
#[serde(alias = "deploymentID", default)]
pub deployment_id: String,
#[serde(default)]
@@ -531,6 +537,85 @@ mod tests {
assert_eq!(value["totalDowntime"], 90);
}
#[test]
fn bucket_target_persisted_wire_keys_stay_snake_case() {
// bucket-targets.json (persisted via `serde_json::to_vec(&BucketTargets)`
// in the admin set/remove handlers) and the msgpack struct-map form
// (`BucketTargets::marshal_msg`) both come straight from this struct's
// serde field names. madmin naming is applied only in the admin
// response layer (`remote_target_admin_json`); renaming here would
// silently break every existing deployment's persisted metadata.
let targets = BucketTargets {
targets: vec![BucketTarget {
credentials: Some(Credentials {
access_key: "ak".to_string(),
secret_key: "sk".to_string(),
session_token: Some("token".to_string()),
expiration: None,
}),
bandwidth_limit: 5,
storage_class: "STANDARD".to_string(),
reset_id: "reset-1".to_string(),
deployment_id: "deploy-1".to_string(),
..Default::default()
}],
};
let json = serde_json::to_value(&targets).expect("targets should serialize to JSON");
let msgpack: serde_json::Value =
rmp_serde::from_slice(&targets.marshal_msg().expect("targets should marshal to msgpack"))
.expect("msgpack struct map should decode into a JSON value");
for (wire, entry) in [("JSON", &json["targets"][0]), ("msgpack", &msgpack["targets"][0])] {
assert_eq!(entry["bandwidth_limit"], 5, "{wire} key `bandwidth_limit` must stay");
assert_eq!(entry["storage_class"], "STANDARD", "{wire} key `storage_class` must stay");
assert_eq!(entry["reset_id"], "reset-1", "{wire} key `reset_id` must stay");
assert_eq!(entry["deployment_id"], "deploy-1", "{wire} key `deployment_id` must stay");
assert_eq!(entry["credentials"]["session_token"], "token", "{wire} key `session_token` must stay");
}
}
#[test]
fn minio_written_bucket_targets_json_populates_madmin_named_fields() {
// A MinIO-written bucket-targets.json carries madmin's JSON tags
// (`bandwidthlimit`, `storageclass`, `resetID`, `deploymentID`,
// `credentials.sessionToken` — madmin-go v3.0.109 bucket-targets.go).
// On migration these must land in the matching fields instead of
// silently defaulting (backlog#1951).
let targets: BucketTargets = serde_json::from_value(serde_json::json!({
"targets": [{
"sourcebucket": "src",
"endpoint": "minio.example:9000",
"credentials": {
"accessKey": "ak",
"secretKey": "sk",
"sessionToken": "minio-session-token"
},
"targetbucket": "dst",
"type": "replication",
"replicationSync": true,
"bandwidthlimit": 107374182400i64,
"storageclass": "STANDARD",
"resetID": "reset-789",
"deploymentID": "deploy-123"
}]
}))
.expect("MinIO-written bucket-targets.json must deserialize");
let target = &targets.targets[0];
assert_eq!(target.bandwidth_limit, 107374182400);
assert_eq!(target.storage_class, "STANDARD");
assert_eq!(target.reset_id, "reset-789");
assert_eq!(target.deployment_id, "deploy-123");
assert_eq!(
target
.credentials
.as_ref()
.and_then(|credentials| credentials.session_token.as_deref()),
Some("minio-session-token")
);
}
#[test]
fn test_bucket_target_debug_redacts_credentials() {
let target = BucketTarget {
+1 -14
View File
@@ -2124,26 +2124,13 @@ impl SetDisks {
let put_object_size = known_put_object_storage_size(data.size());
let shard_file_size_raw = erasure.shard_file_size(put_object_size);
let is_inline_buffer =
storage_class_config.should_inline(shard_file_size_raw, erasure.data_shards, opts.versioned);
let is_inline_buffer = storage_class_config.should_inline(shard_file_size_raw, erasure.data_shards, opts.versioned);
let collect_stage_timing = rustfs_io_metrics::put_stage_metrics_enabled() || issue3031_diag_enabled();
let shard_file_size = shard_file_size_raw;
let shard_size = erasure.shard_size();
let write_path = classify_put_write_path(is_inline_buffer, put_object_size, fi.erasure.block_size);
let direct_inline_commit = matches!(write_path, SmallWritePath::Inline);
{
use std::io::Write;
let msg = format!(
"INLINE_DEBUG: bucket={} obj={} size={} shard_fs={} ds={} bs={} inline={} direct={} path={} iblock={} ver={}\n",
bucket, object, put_object_size, shard_file_size_raw, erasure.data_shards, fi.erasure.block_size,
is_inline_buffer, direct_inline_commit, write_path.metric_label(), storage_class_config.inline_block(), opts.versioned
);
if let Ok(mut f) = std::fs::OpenOptions::new().create(true).append(true).open("/tmp/rustfs_inline_debug.log") {
let _ = f.write_all(msg.as_bytes());
}
let _ = std::io::stderr().write_all(msg.as_bytes());
}
rustfs_io_metrics::record_put_object_path(write_path.metric_label());
let writer_setup_stage_start = collect_stage_timing.then(Instant::now);
let (mut writers, errors) = if direct_inline_commit {
+1 -1
View File
@@ -3194,7 +3194,7 @@ impl ECStore {
// Default return value
let mut del_objects = vec![DeletedObject::default(); objects.len()];
let mut accounting = vec![None; objects.len()];
let accounting = vec![None; objects.len()];
let mut del_errs = Vec::with_capacity(objects.len());
for _ in 0..objects.len() {
@@ -271,7 +271,7 @@ pub(super) fn resolve_latest_object_info_candidates(
.filter(|candidate| latest_candidate_mod_time(candidate) == Some(latest_mod_time))
.collect::<Vec<_>>();
latest_candidates.sort_by(|left, right| right.idx.cmp(&left.idx));
latest_candidates.sort_by_key(|candidate| std::cmp::Reverse(candidate.idx));
let Some(winner) = latest_candidates.first() else {
return Err(Error::ErasureReadQuorum);
-3
View File
@@ -43,9 +43,6 @@ allow-git = [
# RustFS fork carrying presigned expiry and constant-time authentication fixes.
# owner: rustfs-maintainers review: 2026-10
"https://github.com/rustfs/s3s.git",
# MiMalloc fork pinned for hotpath allocation counting support.
# owner: houseme review: 2026-10
"https://github.com/xonatius/mimalloc_rust.git",
]
[bans]
+149
View File
@@ -0,0 +1,149 @@
# CI gate matrix
This file is the source of truth for which validation runs on each event, its
configured wall-clock budget, and whether it can block a merge. Test taxonomy,
naming, and nextest serialization rules remain in [README.md](README.md); e2e
membership and counts remain in
[e2e-suite-inventory.md](e2e-suite-inventory.md).
The distinction between **required** and **report-only** is load-bearing:
a failing job blocks a merge only when its exact check name is present in the
live `main` ruleset. A workflow name, a `merge_group` trigger, or a red PR check
does not make a job required by itself.
## Required merge checks
The live `main` ruleset (`6436880`) currently requires exactly these contexts:
| Required context | Producer | Validation |
|---|---|---|
| `CLA Check` | `.github/workflows/cla.yml` | Contributor agreement |
| `Quick Checks` | `.github/workflows/ci.yml` | Formatting and repository guard scripts |
| `Test and Lint` | `.github/workflows/ci.yml` | Clippy, workspace nextest excluding `e2e_test`, doctests, and migration proofs |
For pull requests limited to the paths excluded by the main CI workflow,
`.github/workflows/ci-docs-only.yml` reports `Quick Checks` and
`Test and Lint` under the same names. It runs the real quick checks and the
planning-document guard; it does not claim that Rust compilation or runtime
tests ran. Despite the workflow name, these paths also include selected deploy,
workflow, and lock files.
Verify the live rule rather than trusting this snapshot before changing merge
policy:
```bash
gh api repos/rustfs/rustfs/rulesets/6436880 \
--jq '.rules[] | select(.type == "required_status_checks") | .parameters'
```
The ruleset currently has `strict_required_status_checks_policy=false`.
`Continuous Integration` accepts `merge_group` events and runs `e2e-full` for
them, but `End-to-End Tests (full merge gate)` is not currently a required
context. Therefore the repository is prepared to test a merge-queue SHA, but
the workflow alone does not prove that every merge passed that lane.
## Pull request and merge matrix
Budgets below are job `timeout-minutes`, not typical runtimes. “Report-only”
means the result is visible and actionable but is not in the live required
context list.
| Event | Validation | Budget | Merge status | Reproduction |
|---|---|---:|---|---|
| PR, non-doc change | `Quick Checks` | 10 min | Required | `make pre-commit` (broader local umbrella) |
| PR, non-doc change | `Test and Lint` | 90 min | Required | `cargo nextest run --profile ci --all --exclude e2e_test` |
| PR, non-doc change | `Typos` | 10 min | Report-only | `typos` |
| PR, non-doc change | `ILM Integration (serial)` | 90 min | Report-only | Use the exact command in `.github/workflows/ci.yml` |
| PR, non-doc change | rio-v2 / swift / sftp test-and-lint variants | 90 min each | Report-only | `cargo nextest run` with the workflow's feature set |
| PR, non-doc change | `Build RustFS Debug Binary` | 30 min | Report-only; prerequisite for black-box lanes | `cargo build -p rustfs --bins` |
| PR, non-doc change | `io_uring Integration (real)` | 30 min | Report-only | `cargo test -p rustfs-ecstore --lib uring_ -- --test-threads=1 --nocapture` |
| PR, non-doc change | `End-to-End Tests` (`e2e-smoke` plus `s3s-e2e`) | 30 min | Report-only | `cargo nextest run --profile e2e-smoke -p e2e_test`; then `./scripts/e2e-run.sh ./target/debug/rustfs <data-dir>` |
| PR, non-doc change | `S3 Implemented Tests` | 60 min | Report-only | Build `rustfs`, then run `scripts/s3-tests/run.sh` with `DEPLOY_MODE=binary`, `TEST_MODE=single`, and `MAXFAIL=0` |
| PR, non-doc change | `S3 Lifecycle Behavior Tests` | 30 min | Report-only | Use the accelerated scanner environment in `.github/workflows/ci.yml` with `scripts/s3-tests/run.sh` |
| PR touching dependency or workflow inputs | Cargo Deny / Workflow Pin Report / Dependency Review | 20 / 5 / 30 min | Report-only | `cargo deny check`; `scripts/security/check_workflow_pins.sh` |
| PR touching architecture rules or architecture docs | `Architecture Migration Rules` | 10 min | Report-only | `scripts/check_architecture_migration_rules.sh` |
| PR touching Nix or workspace manifests | `Nix Build & Check` | 60 min | Report-only | `nix flake check` |
| PR limited to main-CI-excluded paths | companion `Quick Checks` and `Test and Lint` | 10 min each | Required | `git diff --check`; `make doc-paths-check` when documentation paths changed |
| `merge_group` | Standard CI plus `e2e-full` | 55 min for `e2e-full` | Standard required contexts only; `e2e-full` report-only | `cargo nextest run --profile e2e-full -p e2e_test` |
| Push to `main` | Standard CI plus `e2e-full` | 55 min for `e2e-full` | Post-merge detection | Same as `merge_group` |
| PR touching fuzz inputs or harness paths | Build plus five 60-second fuzz smoke targets | 60 min build; 30 min per target | Report-only | `MAX_TOTAL_TIME=60 ./scripts/fuzz/run.sh` |
| PR touching selected ecstore disk/format paths | `Rename Safety` on Windows | 60 min | Report-only | Run the four `cargo test -p rustfs-ecstore --lib <filter>` commands in `windows-filesystem.yml` on Windows |
The authoritative e2e filters live in `.config/nextest.toml`; extend a profile
instead of adding a second ad-hoc selector. Before a profile runs,
`scripts/check_test_wiring.py` compares its exact membership to the committed
digest so a silent test drop fails closed.
## Scheduled and manual validation
Scheduled lanes are independent fault domains. They do not block a pull
request, but their workflow-local gate can fail the run and scheduled failures
are routed to the shared failure-issue action. The scheduled-validation
watchdog and freshness workflow separately detect incomplete runs and missing
schedules.
| Cadence (UTC unless noted) | Workflow / validation | Budget | Verdict and artifacts | Reproduction |
|---|---|---:|---|---|
| Daily 02:17 | Fuzz: five nightly corpus targets | 60 min build; 60 min per target | Gate; corpus/crash artifacts, scheduled failure alert | `MAX_TOTAL_TIME=<seconds> ./scripts/fuzz/run.sh` |
| Daily 03:17 | MinIO interop (EC + SSE read parity) | 40 min | Gate; scheduled failure alert | Dispatch `minio-interop.yml` or follow its pinned Docker fixture steps |
| Daily 04:29 | Replication / cluster-fault / protocol e2e | 45 / 90 / 90 min | Three independent gates; JUnit, membership, and server logs | `cargo nextest run --profile e2e-repl-nightly -p e2e_test`; `--profile e2e-nightly`; `-j 1 --profile e2e-protocols` |
| Daily 06:31 | Warp performance A/B | 180 min | Regression budget gate; A/B summaries and server logs | `bash scripts/run_hotpath_warp_abba.sh --help` |
| Daily 00:07 Asia/Shanghai (16:07 UTC previous day) | Nightly GNU build and Vault lanes | 150 / 90 / 60 min | Build, live Vault, and HA failover gates | Use the commands and pinned Vault images in `nightly-gnu.yml` |
| Daily 03:23 | Security Audit | 20 / 5 min, plus 30 min on PR dependency review | Cargo Deny and workflow-pin gates; scheduled failure alert | `cargo deny check`; `scripts/security/check_workflow_pins.sh` |
| Daily 23:47 | Scheduled Validation Freshness | 10 min | Fails when a critical schedule was never created or is stale | Dispatch `scheduled-validation-freshness.yml` |
| Sunday 00:11 | Full `Continuous Integration` matrix | Per-job budgets above | Weekly variant coverage, including dormant rio-v2 binary/e2e lanes | Dispatch `ci.yml` |
| Sunday 01:13 | Seven-platform build matrix | 150 min per platform | Build/package integrity; scheduled failure alert | Dispatch `build.yml` with an exact platform set |
| Sunday 02:19 | Ceph s3-tests full sweep: single and real four-node, four shards each | 180 min per shard | Compatibility gate; report, JUnit, exact node IDs, and server logs | `scripts/s3-tests/run.sh` against an existing single or distributed target |
| Sunday 06:41 | Mint | 120 min | **Report-only by design**; per-suite PASS/FAIL/NA and raw `log.json` | Reproduce the pinned Docker sequence in `mint.yml` or dispatch it |
| Sunday 07:43 | Workspace line coverage | 120 min | Report-only trend; lcov and JSON retained 90 days | `make coverage` |
| Monthly, day 1 06:37 | Runner Hygiene | 15 min | Validates runner ephemerality; scheduled failure alert | Dispatch `runner-hygiene.yml` |
Manual `workflow_dispatch` exists for the scheduled workflows above. Manual
runs are debugging evidence and intentionally do not open scheduled-failure
issues. A manual performance run may explicitly allow a known regression; that
override must not be treated as an ordinary passing baseline.
## Release validation
Release validation is post-merge and tag-driven; it does not substitute for a
pull-request gate.
| Event | Validation | Budget | Result |
|---|---|---:|---|
| Push to `main` or weekly schedule | `Build and Release` platform matrix | 150 min per platform | Build artifacts for all selected targets; no release publication on a main push |
| Valid release or preview tag | `Build and Release` plus asset checks | 150 min per platform | Draft release, checksummed assets, and publish step |
| Successful non-preview release-tag build | Docker image build and image scan | 60 min build; 30 min scan | Multi-architecture images plus vulnerability report |
| Successful release-tag build | DEB/RPM packaging | 30 min per architecture | Packages and checksum files uploaded to the release |
| Successful non-preview release-tag build | Helm template test and package | 30 min build; 30 min publish | Versioned chart and repository index |
Use an exact preview tag for end-to-end release rehearsal. Manual dispatches
are backfill/debug paths and do not prove the automatic `workflow_run` chain.
## Evidence requirements
A green check is useful only when it proves the intended behavior ran:
- Record the exact commit SHA and run URL.
- Separate product failure from runner prerequisites, service readiness, and
cancellation. Repair the precondition, then rerun the exact workload.
- Preserve membership manifests, JUnit, raw compatibility logs, seeds, and
server logs where the workflow provides them.
- For a bug fix or a new fault checker, provide sensitivity evidence: the old
behavior or an intentional mutation must fail the new oracle, and the fixed
behavior must pass it.
- Never promote a report-only lane to required from one green run. Require at
least 14 days and 30 representative pull requests with at least 99% complete
execution, then update the ruleset and this table together.
## Change checklist
Update this file in the same pull request when any of these change:
- workflow triggers, job names, timeouts, or nextest profile ownership;
- required status contexts or strict/merge-queue policy;
- scheduled cadence, alert routing, artifact contract, or local reproduction;
- report-only versus gating semantics.
Do not copy per-module test counts here. Update
[e2e-suite-inventory.md](e2e-suite-inventory.md) and its enforced membership
digest instead.
+2 -2
View File
@@ -336,13 +336,13 @@ opentelemetry = { workspace = true }
tracing-opentelemetry = { workspace = true }
# Data structures
hashbrown = { workspace = true, features = ["serde", "rayon"] }
mimalloc = { workspace = true }
rustfs-mimalloc = { workspace = true }
[target.'cfg(target_os = "linux")'.dependencies]
libsystemd.workspace = true
[target.'cfg(not(target_os = "windows"))'.dependencies]
libmimalloc-sys.workspace = true
rustfs-mimalloc-sys.workspace = true
[dev-dependencies]
uuid = { workspace = true, features = ["v4", "v5", "fast-rng", "macro-diagnostics"] }
+139 -2
View File
@@ -376,7 +376,10 @@ impl RemoteTargetRequest {
/// Admin-response encoding of a remote target: the persisted bucket-targets
/// format keeps `healthCheckDuration`/`totalDowntime` in seconds and the
/// `latency` stats in milliseconds, but madmin decodes all of them as Go
/// `time.Duration` (nanoseconds) — re-encode just those fields without
/// `time.Duration` (nanoseconds) — and it looks the fields up under its own
/// JSON tags (`bandwidthlimit`, `storageclass`, `resetID`, `deploymentID`,
/// `credentials.sessionToken` — madmin-go v3.0.109 `bucket-targets.go`), not
/// the persisted snake_case keys. Re-encode just those fields here without
/// touching the persistence wire format.
fn remote_target_admin_json(target: &BucketTarget) -> Result<serde_json::Value, serde_json::Error> {
fn go_duration_nanos(duration: Duration) -> serde_json::Value {
@@ -385,6 +388,12 @@ fn remote_target_admin_json(target: &BucketTarget) -> Result<serde_json::Value,
u64::try_from(duration.as_nanos()).unwrap_or(u64::MAX).into()
}
fn rename_key(value: &mut serde_json::Value, from: &str, to: &str) {
if let Some(moved) = value.as_object_mut().and_then(|object| object.remove(from)) {
value[to] = moved;
}
}
let mut value = serde_json::to_value(target)?;
value["healthCheckDuration"] = go_duration_nanos(target.health_check_duration);
value["totalDowntime"] = go_duration_nanos(target.total_downtime);
@@ -393,6 +402,11 @@ fn remote_target_admin_json(target: &BucketTarget) -> Result<serde_json::Value,
"avg": go_duration_nanos(target.latency.avg),
"max": go_duration_nanos(target.latency.max),
});
rename_key(&mut value, "bandwidth_limit", "bandwidthlimit");
rename_key(&mut value, "storage_class", "storageclass");
rename_key(&mut value, "reset_id", "resetID");
rename_key(&mut value, "deployment_id", "deploymentID");
rename_key(&mut value["credentials"], "session_token", "sessionToken");
Ok(value)
}
@@ -1476,7 +1490,7 @@ mod tests {
parse_remote_target_update_ops, render_mrf_backlog, render_replication_diff, unique_replication_peers,
validate_remote_target_tls_settings,
};
use crate::admin::storage_api::bucket::target::{BucketTarget, LatencyStat};
use crate::admin::storage_api::bucket::target::{BucketTarget, Credentials as TargetCredentials, LatencyStat};
use crate::admin::storage_api::replication::{BucketStats, DurableMrfBacklog, MrfOpKind, MrfReplicateEntry};
use http::Uri;
@@ -2272,6 +2286,129 @@ mod tests {
assert_eq!(persisted["latency"]["max"], 250);
}
#[test]
fn list_remote_targets_response_uses_madmin_key_names() {
// madmin-go v3.0.109 BucketTarget JSON tags are `bandwidthlimit`,
// `storageclass`, `resetID`, `deploymentID`, and
// `credentials.sessionToken` (backlog#1951); the persisted snake_case
// keys decode to zero values in mc, blanking the bandwidth and
// reset-id columns of `mc replicate ls`.
let target = BucketTarget {
endpoint: "192.168.1.10:9000".to_string(),
target_bucket: "target".to_string(),
credentials: Some(TargetCredentials {
access_key: "access".to_string(),
secret_key: String::new(),
session_token: Some("session-token".to_string()),
expiration: None,
}),
bandwidth_limit: 107_374_182_400,
storage_class: "STANDARD".to_string(),
reset_id: "reset-123".to_string(),
deployment_id: "deploy-456".to_string(),
..Default::default()
};
let value = super::remote_target_admin_json(&target).expect("admin response should serialize");
assert_eq!(value["bandwidthlimit"], 107_374_182_400i64);
assert_eq!(value["storageclass"], "STANDARD");
assert_eq!(value["resetID"], "reset-123");
assert_eq!(value["deploymentID"], "deploy-456");
assert_eq!(value["credentials"]["sessionToken"], "session-token");
// The madmin keys replace the snake_case ones rather than duplicating
// them next to each other.
for stale in ["bandwidth_limit", "bandwidth", "storage_class", "reset_id", "deployment_id"] {
assert!(value.get(stale).is_none(), "admin response must not carry `{stale}`");
}
assert!(value["credentials"].get("session_token").is_none());
}
/// Decode-side mirror of madmin-go v3.0.109 `BucketTarget`/`Credentials`
/// (`bucket-targets.go`): the exact `json:"..."` tags mc's `encoding/json`
/// looks fields up under. Unknown keys are ignored like Go does, and a
/// missing key leaves the Go zero value, which is exactly how a misnamed
/// key turns into a blank column in `mc replicate ls`.
#[derive(Debug, Default, serde::Deserialize)]
#[serde(default)]
struct MadminBucketTarget {
sourcebucket: String,
endpoint: String,
credentials: Option<MadminCredentials>,
targetbucket: String,
arn: String,
bandwidthlimit: i64,
#[serde(rename = "replicationSync")]
replication_sync: bool,
storageclass: String,
#[serde(rename = "healthCheckDuration")]
health_check_duration: i64,
#[serde(rename = "resetID")]
reset_id: String,
#[serde(rename = "totalDowntime")]
total_downtime: i64,
#[serde(rename = "deploymentID")]
deployment_id: String,
}
#[derive(Debug, Default, serde::Deserialize)]
#[serde(default)]
struct MadminCredentials {
#[serde(rename = "accessKey")]
access_key: String,
#[serde(rename = "secretKey")]
secret_key: String,
#[serde(rename = "sessionToken")]
session_token: String,
}
#[test]
fn list_remote_targets_response_decodes_through_madmin_tags() {
// Regression for the review on backlog#1951: the response must decode
// a nonzero bandwidth limit through madmin's `bandwidthlimit` tag (not
// `bandwidth`, which Go would silently drop as an unknown key).
let target = BucketTarget {
source_bucket: "src".to_string(),
endpoint: "192.168.1.10:9000".to_string(),
target_bucket: "target".to_string(),
arn: "arn:rustfs:replication:us-east-1:dep:target".to_string(),
credentials: Some(TargetCredentials {
access_key: "access".to_string(),
secret_key: String::new(),
session_token: Some("session-token".to_string()),
expiration: None,
}),
bandwidth_limit: 1_073_741_824,
replication_sync: true,
storage_class: "STANDARD".to_string(),
health_check_duration: std::time::Duration::from_secs(60),
reset_id: "reset-123".to_string(),
total_downtime: std::time::Duration::from_secs(90),
deployment_id: "deploy-456".to_string(),
..Default::default()
};
let wire = serde_json::to_string(&super::remote_target_admin_json(&target).expect("admin response should serialize"))
.expect("admin response should encode");
let decoded: MadminBucketTarget = serde_json::from_str(&wire).expect("madmin-shaped decode must succeed");
assert_eq!(decoded.bandwidthlimit, 1_073_741_824, "mc must see the nonzero bandwidth limit");
assert_eq!(decoded.sourcebucket, "src");
assert_eq!(decoded.endpoint, "192.168.1.10:9000");
assert_eq!(decoded.targetbucket, "target");
assert_eq!(decoded.arn, "arn:rustfs:replication:us-east-1:dep:target");
assert!(decoded.replication_sync);
assert_eq!(decoded.storageclass, "STANDARD");
assert_eq!(decoded.health_check_duration, 60_000_000_000);
assert_eq!(decoded.reset_id, "reset-123");
assert_eq!(decoded.total_downtime, 90_000_000_000);
assert_eq!(decoded.deployment_id, "deploy-456");
let credentials = decoded.credentials.expect("credentials must decode");
assert_eq!(credentials.access_key, "access");
assert_eq!(credentials.secret_key, "");
assert_eq!(credentials.session_token, "session-token");
}
#[test]
fn remote_target_admin_json_latency_round_trips_through_go_duration() {
// Round trip: a madmin reader decodes the latency values as Go
+1 -7
View File
@@ -369,14 +369,8 @@ pub fn allocator_reclaim_controller_snapshot(ctx: &CancellationToken) -> Allocat
}
#[cfg(not(target_os = "windows"))]
#[allow(unsafe_code)]
fn collect_allocator_memory(force: bool) -> Result<(), String> {
// SAFETY: `mi_collect` is provided by the active global allocator backend
// on this target family. It is explicitly intended to reclaim retained
// pages/segments and does not require additional invariants from the caller.
unsafe {
libmimalloc_sys::mi_collect(force);
}
rustfs_mimalloc::MiMalloc::collect(force);
Ok(())
}
+10 -8
View File
@@ -26,22 +26,22 @@ struct MiMallocAllocator;
unsafe impl GlobalAlloc for MiMallocAllocator {
unsafe fn alloc(&self, layout: Layout) -> *mut u8 {
// SAFETY: the caller upholds GlobalAlloc's contract for layout.
unsafe { mimalloc::MiMalloc.alloc(layout) }
unsafe { rustfs_mimalloc::MiMalloc.alloc(layout) }
}
unsafe fn alloc_zeroed(&self, layout: Layout) -> *mut u8 {
// SAFETY: the caller upholds GlobalAlloc's contract for layout.
unsafe { mimalloc::MiMalloc.alloc_zeroed(layout) }
unsafe { rustfs_mimalloc::MiMalloc.alloc_zeroed(layout) }
}
unsafe fn dealloc(&self, ptr: *mut u8, layout: Layout) {
// SAFETY: ptr and layout came from this allocator and are forwarded unchanged.
unsafe { mimalloc::MiMalloc.dealloc(ptr, layout) }
unsafe { rustfs_mimalloc::MiMalloc.dealloc(ptr, layout) }
}
unsafe fn realloc(&self, ptr: *mut u8, layout: Layout, new_size: usize) -> *mut u8 {
// SAFETY: ptr and layout came from this allocator and are forwarded unchanged.
unsafe { mimalloc::MiMalloc.realloc(ptr, layout, new_size) }
unsafe { rustfs_mimalloc::MiMalloc.realloc(ptr, layout, new_size) }
}
}
@@ -51,7 +51,7 @@ static GLOBAL: hotpath::CountingAllocator<MiMallocAllocator> = hotpath::Counting
#[cfg(not(all(feature = "hotpath", feature = "hotpath-alloc")))]
#[global_allocator]
static GLOBAL: mimalloc::MiMalloc = mimalloc::MiMalloc;
static GLOBAL: rustfs_mimalloc::MiMalloc = rustfs_mimalloc::MiMalloc;
fn main() {
let _hotpath_guard = hotpath::HotpathGuardBuilder::new("main").build();
@@ -71,8 +71,9 @@ mod tests {
allocation.extend_from_slice(&[7_u8; 64]);
assert_eq!(allocation.len(), 64);
let heap = rustfs_mimalloc::heap::Heap::main();
// SAFETY: the live Vec pointer is valid to inspect for heap ownership.
assert!(unsafe { libmimalloc_sys::mi_is_in_heap_region(allocation.as_ptr().cast()) });
assert!(unsafe { heap.contains(allocation.as_ptr()) });
}
#[test]
@@ -85,12 +86,13 @@ mod tests {
let layout = Layout::from_size_align(32, 8).expect("valid test allocation layout");
let grown_layout = Layout::from_size_align(64, 8).expect("valid grown test allocation layout");
let allocator = super::MiMallocAllocator;
let heap = rustfs_mimalloc::heap::Heap::main();
// SAFETY: The pointer is checked for null before use and later released
// through the same allocator with the corresponding layout.
let ptr = unsafe { allocator.alloc_zeroed(layout) };
assert!(!ptr.is_null());
assert!(unsafe { libmimalloc_sys::mi_is_in_heap_region(ptr.cast()) });
assert!(unsafe { heap.contains(ptr) });
assert!(unsafe { std::slice::from_raw_parts(ptr, 32).iter().all(|byte| *byte == 0) });
// SAFETY: `ptr` was allocated by `allocator` with `layout`; on failure
@@ -102,7 +104,7 @@ mod tests {
panic!("mimalloc realloc failed in allocator smoke test");
}
assert!(unsafe { libmimalloc_sys::mi_is_in_heap_region(grown_ptr.cast()) });
assert!(unsafe { heap.contains(grown_ptr) });
// SAFETY: `grown_ptr` was reallocated by `allocator` and is released
// with the matching grown layout.
unsafe { allocator.dealloc(grown_ptr, grown_layout) };
+19 -36
View File
@@ -17,10 +17,7 @@ use rustfs_io_metrics::{
record_cpu_usage, record_memory_usage, record_process_memory_split,
};
use serde::Serialize;
#[cfg(any(test, not(target_os = "windows")))]
use serde_json::Value;
#[cfg(not(target_os = "windows"))]
use std::ffi::CStr;
use std::path::Path;
use std::sync::{Arc, Mutex, OnceLock};
use std::time::Duration;
@@ -231,7 +228,18 @@ fn read_cgroup_memory_snapshot() -> Option<CgroupMemorySnapshot> {
read_cgroup_v2().or_else(read_cgroup_v1)
}
#[cfg(any(test, not(target_os = "windows")))]
fn read_allocator_memory_snapshot() -> Option<AllocatorMemorySnapshot> {
let json = rustfs_mimalloc::MiMalloc::stats_json();
if json.is_empty() {
return None;
}
let observation = parse_mimalloc_stats_json(&json)?;
Some(AllocatorMemorySnapshot {
backend: crate::allocator_reclaim::allocator_backend(),
observation,
})
}
fn numeric_json_value(value: &Value) -> Option<u64> {
match value {
Value::Number(number) => number
@@ -242,7 +250,6 @@ fn numeric_json_value(value: &Value) -> Option<u64> {
}
}
#[cfg(any(test, not(target_os = "windows")))]
fn numeric_json_field(value: &Value, field: &str) -> Option<u64> {
match value {
Value::Object(fields) => fields
@@ -254,7 +261,6 @@ fn numeric_json_field(value: &Value, field: &str) -> Option<u64> {
}
}
#[cfg(any(test, not(target_os = "windows")))]
fn mimalloc_stat_field(value: &Value, metric: &str, field: &str) -> Option<u64> {
match value {
Value::Object(fields) => {
@@ -271,12 +277,10 @@ fn mimalloc_stat_field(value: &Value, metric: &str, field: &str) -> Option<u64>
}
}
#[cfg(any(test, not(target_os = "windows")))]
fn mimalloc_stat_current(value: &Value, metric: &str) -> Option<u64> {
mimalloc_stat_field(value, metric, "current")
}
#[cfg(any(test, not(target_os = "windows")))]
fn mimalloc_stat_sum(value: &Value, metrics: &[&str], field: &str) -> Option<u64> {
metrics
.iter()
@@ -285,7 +289,6 @@ fn mimalloc_stat_sum(value: &Value, metrics: &[&str], field: &str) -> Option<u64
.filter(|value| *value > 0)
}
#[cfg(any(test, not(target_os = "windows")))]
fn parse_mimalloc_stats_json(stats_json: &str) -> Option<AllocatorMemoryObservation> {
let value = serde_json::from_str::<Value>(stats_json).ok()?;
let malloc_metrics = ["malloc_normal", "malloc_huge"];
@@ -312,33 +315,6 @@ fn parse_mimalloc_stats_json(stats_json: &str) -> Option<AllocatorMemoryObservat
}
}
#[cfg(not(target_os = "windows"))]
#[allow(unsafe_code)]
fn read_allocator_memory_snapshot() -> Option<AllocatorMemorySnapshot> {
// SAFETY: `mi_stats_get_json` returns a null-terminated JSON buffer owned by
// mimalloc when called with a null input buffer. The mimalloc API requires
// freeing that buffer with `mi_free`; parsing finishes before the buffer is freed.
let observation = unsafe {
let stats_ptr = libmimalloc_sys::mi_stats_get_json(0, std::ptr::null_mut());
if stats_ptr.is_null() {
return None;
}
let observation = CStr::from_ptr(stats_ptr).to_str().ok().and_then(parse_mimalloc_stats_json);
libmimalloc_sys::mi_free(stats_ptr.cast());
observation?
};
Some(AllocatorMemorySnapshot {
backend: crate::allocator_reclaim::allocator_backend(),
observation,
})
}
#[cfg(target_os = "windows")]
fn read_allocator_memory_snapshot() -> Option<AllocatorMemorySnapshot> {
None
}
fn configured_memory_observability_interval_secs() -> u64 {
rustfs_utils::get_env_u64(ENV_MEMORY_OBSERVABILITY_INTERVAL_SECS, DEFAULT_MEMORY_OBSERVABILITY_INTERVAL_SECS).max(1)
}
@@ -566,6 +542,13 @@ mod tests {
assert_eq!(parse_mimalloc_stats_json(r#"{ "allocator": "unknown" }"#), None);
}
#[test]
fn read_allocator_memory_snapshot_uses_mimalloc_stats_json() {
let snapshot = super::read_allocator_memory_snapshot();
#[cfg(not(target_os = "windows"))]
assert!(snapshot.is_some(), "allocator snapshot should be available on non-Windows");
}
#[test]
fn memory_observability_snapshot_reports_disabled_when_metrics_are_disabled() {
let snapshot = build_memory_observability_status_snapshot(false, 15, false);
+26 -11
View File
@@ -34,8 +34,8 @@ CONCURRENCY=8
DURATION="60s"
ROUNDS=3
COOLDOWN_SECS=20
DATASET_SETUP_DURATION="10s"
HEALTH_TIMEOUT_SECS=180
DATASET_OBJECTS_PER_WORKER=8
FAIL_PCT=10
WARN_PCT=5
ALLOW_REGRESSION=false
@@ -100,6 +100,9 @@ Benchmark:
--duration <dur> warp duration per cell (default 60s).
--rounds <n> rounds per cell; must be >= 3 (default 3).
--cooldown <n> cooldown seconds between rounds/sizes (default 20).
--dataset-setup-duration <dur>
isolated Warp PUT warm-up for get/mixed legs
(default 10s; not included in the measurement).
--concurrency <n> warp concurrency (default 8).
--warp-bin <path> warp binary (default warp).
@@ -170,6 +173,7 @@ while [[ $# -gt 0 ]]; do
--duration) DURATION="$2"; shift 2 ;;
--rounds) ROUNDS="$2"; shift 2 ;;
--cooldown) COOLDOWN_SECS="$2"; shift 2 ;;
--dataset-setup-duration) DATASET_SETUP_DURATION="$2"; shift 2 ;;
--health-timeout) HEALTH_TIMEOUT_SECS="$2"; shift 2 ;;
--fail-pct) FAIL_PCT="$2"; shift 2 ;;
--warn-pct) WARN_PCT="$2"; shift 2 ;;
@@ -362,18 +366,29 @@ measure() {
--duration "$DURATION" --rounds "$ROUNDS" --cooldown-secs "$COOLDOWN_SECS"
--out-dir "$cell"
)
if [[ "$mode" != "put" ]]; then
# Warp defaults to 2,500 setup objects per round. At 10 MiB that writes
# 25 GiB before every 12-second measurement, so the matrix cannot finish
# inside the workflow budget. Eight objects per worker keeps preparation
# bounded while retaining a multi-object working set for relative A/B.
args+=(--extra-args "--objects $((CONCURRENCY * DATASET_OBJECTS_PER_WORKER)) --noclear")
fi
[[ "$mode" == "put" ]] || args+=(--extra-args "--noclear")
[[ -n "$baseline_csv" ]] && args+=(--baseline-csv "$baseline_csv")
run "$ENHANCED_BENCH" "${args[@]}" >&2
echo "$cell"
}
prepare_dataset() {
local leg="$1" workload="$2" mode="$3" size="$4" sync_label="$5" bucket="$6"
[[ "$mode" != "put" ]] || return 0
local setup_cell="$OUT_DIR/$workload/$sync_label/$leg/dataset-setup"
local args=(
--tool warp --warp-bin "$WARP_BIN" --warp-mode put
--endpoint "$ADDRESS" --access-key "$ACCESS_KEY" --secret-key "$SECRET_KEY"
--region "$REGION" --bucket "$bucket" --sizes "$size" --concurrency "$CONCURRENCY"
--duration "$DATASET_SETUP_DURATION" --rounds 1 --cooldown-secs 0
--extra-args "--noclear"
--out-dir "$setup_cell"
)
log "preparing isolated dataset: $sync_label/$workload/$leg bucket=$bucket"
run "$ENHANCED_BENCH" "${args[@]}" >&2
}
write_schedule_header() {
echo "sync_label,drive_sync,workload,mode,size,leg,phase,binary,out_dir,bucket,dataset_setup" >"$OUT_DIR/abba_schedule.csv"
}
@@ -384,7 +399,7 @@ append_schedule() {
phase="$(phase_for_leg "$leg")"
bin="$(binary_for_leg "$leg")"
local dataset_setup="none"
[[ "$mode" == "put" ]] || dataset_setup="warp-native-bounded"
[[ "$mode" == "put" ]] || dataset_setup="warp-put"
echo "$sync_label,$drive_sync,$workload,$mode,$size,$leg,$phase,$bin,$OUT_DIR/$workload/$sync_label/$leg,$bucket,$dataset_setup" >>"$OUT_DIR/abba_schedule.csv"
}
@@ -466,8 +481,7 @@ dataset_namespace=$DATASET_NAMESPACE
local_run_data_root=$RUN_DATA_ROOT
bucket_isolation=per-leg
bucket_prefix=rustfs-abba-$DATASET_NAMESPACE
dataset_setup=get-and-mixed-via-bounded-warp-native
dataset_objects=$((CONCURRENCY * DATASET_OBJECTS_PER_WORKER))
dataset_setup=get-and-mixed-via-warp-put
endpoint=$ADDRESS
warp_version=$("$WARP_BIN" --version 2>/dev/null | head -n1 || echo unknown)
EOF
@@ -488,6 +502,7 @@ for ds_spec in "${DRIVE_SYNC_MATRIX[@]}"; do
log "=== $sync_label $workload leg $leg ($(phase_for_leg "$leg")) ==="
bucket="$(bucket_for_leg "$sync_label" "$workload" "$leg")"
bring_up "$leg" "$drive_sync" "$workload" "$mode" "$size" "$sync_label" "$bucket"
prepare_dataset "$leg" "$workload" "$mode" "$size" "$sync_label" "$bucket"
append_schedule "$sync_label" "$drive_sync" "$workload" "$mode" "$size" "$leg" "$bucket"
baseline_csv=""
+14 -29
View File
@@ -672,7 +672,7 @@ extract_report_line() {
local regex="$1"
local file="$2"
awk -v regex="$regex" '
/^(Report|Operation):/ {
/^Report:/ {
in_report = 1
next
}
@@ -703,9 +703,9 @@ normalize_duration_metric() {
extract_metrics() {
local log_file="$1"
local average_line request_line throughput reqps latency req_p90 req_p99 reqps_num
local average_line reqs_line throughput reqps latency req_p90 req_p99 reqps_num
average_line="$(extract_report_line '^[[:space:]]*[*][[:space:]]+Average:' "$log_file")"
request_line="$(extract_report_line '^[[:space:]]*[*][[:space:]]+(Reqs:[[:space:]]+)?Avg:' "$log_file")"
reqs_line="$(extract_report_line '^[[:space:]]*[*][[:space:]]+Reqs:' "$log_file")"
if [[ -n "$average_line" ]]; then
throughput="$(echo "$average_line" | sed -E 's/^.*Average:[[:space:]]*//; s/,[[:space:]]*.*$//')"
@@ -715,16 +715,20 @@ extract_metrics() {
reqps="$(extract_first '[0-9]+(\.[0-9]+)?[[:space:]]*(obj/s|req/s|ops/s|requests/s)' "$log_file")"
fi
if [[ -n "$request_line" ]]; then
latency="$(echo "$request_line" | rg -o 'Avg:[[:space:]]+[0-9]+(\.[0-9]+)?(ms|us|µs|s)' | sed -E 's/^Avg:[[:space:]]+//')"
req_p90="$(echo "$request_line" | rg -o '90%:[[:space:]]+[0-9]+(\.[0-9]+)?(ms|us|µs|s)' | sed -E 's/^90%:[[:space:]]+//')"
req_p99="$(echo "$request_line" | rg -o '99%:[[:space:]]+[0-9]+(\.[0-9]+)?(ms|us|µs|s)' | sed -E 's/^99%:[[:space:]]+//')"
if [[ -n "$reqs_line" ]]; then
latency="$(echo "$reqs_line" | rg -o 'Avg:[[:space:]]+[0-9]+(\.[0-9]+)?(ms|us|µs|s)' | sed -E 's/^Avg:[[:space:]]+//')"
req_p90="$(echo "$reqs_line" | rg -o '90%:[[:space:]]+[0-9]+(\.[0-9]+)?(ms|us|µs|s)' | sed -E 's/^90%:[[:space:]]+//')"
req_p99="$(echo "$reqs_line" | rg -o '99%:[[:space:]]+[0-9]+(\.[0-9]+)?(ms|us|µs|s)' | sed -E 's/^99%:[[:space:]]+//')"
else
latency="$(rg -o 'Reqs:[[:space:]]+Avg:[[:space:]]+[0-9]+(\.[0-9]+)?(ms|us|µs|s)' "$log_file" | head -n1 | sed -E 's/^Reqs:[[:space:]]+Avg:[[:space:]]+//')"
req_p90="$(rg -o '90%:[[:space:]]+[0-9]+(\.[0-9]+)?(ms|us|µs|s)' "$log_file" | head -n1 | sed -E 's/^90%:[[:space:]]+//')"
req_p99="$(rg -o '99%:[[:space:]]+[0-9]+(\.[0-9]+)?(ms|us|µs|s)' "$log_file" | head -n1 | sed -E 's/^99%:[[:space:]]+//')"
fi
if [[ -z "$latency" ]]; then
latency="$(extract_first '[0-9]+(\.[0-9]+)?[[:space:]]*(ms|us|µs|s)' "$log_file")"
fi
throughput="$(trim "${throughput:-N/A}")"
reqps="$(trim "${reqps:-N/A}")"
latency="$(trim "${latency:-N/A}")"
@@ -1124,8 +1128,6 @@ run_one_attempt() {
"--concurrent" "$CONCURRENCY"
"--duration" "$DURATION"
"--region" "$REGION"
"--no-color"
"--analyze.v"
)
if [[ "$INSECURE" == "true" ]]; then
cmd+=("--insecure")
@@ -1210,12 +1212,6 @@ run_one_attempt() {
req_p99_ms="$(to_ms "$req_p99_human")"
fi
if [[ "$DRY_RUN" != "true" && "$TOOL" == "warp" && "$status" == "ok" ]] \
&& rg -q '^[[:space:]]*(Total[[:space:]]+)?Errors:[[:space:]]+[1-9][0-9]*[.]?([[:space:]]|$)' "$log_file"; then
status="failed"
exit_code=1
fi
if [[ "$DRY_RUN" != "true" && "$status" == "ok" ]]; then
if [[ "$throughput_bps" == "N/A" && "$reqps" == "N/A" ]]; then
status="failed"
@@ -1322,21 +1318,10 @@ compare_baseline() {
dr="N/A"; dl="N/A"; dt="N/A"; dp90="N/A"; dp99="N/A"; ne="N/A"; be="N/A"; de="N/A"
if (br!="N/A" && n_req!="N/A" && br+0!=0) dr=sprintf("%.2f", ((n_req-br)/br)*100)
if (bl!="N/A" && n_lat!="N/A") {
if (bl+0!=0) dl=sprintf("%.2f", ((n_lat-bl)/bl)*100)
else if (n_lat+0==0) dl="0.00"
}
if (bl!="N/A" && n_lat!="N/A" && bl+0!=0) dl=sprintf("%.2f", ((n_lat-bl)/bl)*100)
if (bt!="N/A" && n_thr!="N/A" && bt+0!=0) dt=sprintf("%.2f", ((n_thr-bt)/bt)*100)
# Warp v1 rounds sub-millisecond latency to 0s. Two zero readings are
# the same below-resolution bucket; a nonzero candidate remains invalid.
if (bp90!="N/A" && n_p90!="N/A") {
if (bp90+0!=0) dp90=sprintf("%.2f", ((n_p90-bp90)/bp90)*100)
else if (n_p90+0==0) dp90="0.00"
}
if (bp99!="N/A" && n_p99!="N/A") {
if (bp99+0!=0) dp99=sprintf("%.2f", ((n_p99-bp99)/bp99)*100)
else if (n_p99+0==0) dp99="0.00"
}
if (bp90!="N/A" && n_p90!="N/A" && bp90+0!=0) dp90=sprintf("%.2f", ((n_p90-bp90)/bp90)*100)
if (bp99!="N/A" && n_p99!="N/A" && bp99+0!=0) dp99=sprintf("%.2f", ((n_p99-bp99)/bp99)*100)
if (n_ok!="N/A" && n_fail!="N/A" && n_ok+n_fail>0) ne=sprintf("%.2f", (n_fail/(n_ok+n_fail))*100)
if (bok!="N/A" && bfail!="N/A" && bok+bfail>0) be=sprintf("%.2f", (bfail/(bok+bfail))*100)
if (ne!="N/A" && be!="N/A") de=sprintf("%.2f", ne-be)
+2 -7
View File
@@ -58,13 +58,8 @@ rg -qx 'evidence_mode=dry-run' "$OUT_DIR/manifest.env"
rg -qx 'formal_evidence=false' "$OUT_DIR/manifest.env"
rg -qx 'performance_conclusion=not_measured_dry_run' "$OUT_DIR/manifest.env"
rg -qx 'bucket_isolation=per-leg' "$OUT_DIR/manifest.env"
rg -qx 'dataset_setup=get-and-mixed-via-bounded-warp-native' "$OUT_DIR/manifest.env"
rg -qx 'dataset_objects=64' "$OUT_DIR/manifest.env"
[[ "$(rg -c -- '--extra-args --objects\\ 64\\ --noclear' "$TRACE_FILE")" == "32" ]]
if rg -q -- 'dataset-setup' "$TRACE_FILE"; then
echo "unexpected redundant dataset setup command" >&2
exit 1
fi
rg -qx 'dataset_setup=get-and-mixed-via-warp-put' "$OUT_DIR/manifest.env"
[[ "$(rg -c -- '--extra-args --noclear' "$TRACE_FILE")" == "64" ]]
! rg -q -- 'rustfs-bench' "$TRACE_FILE"
if "$RUNNER" \
+2 -67
View File
@@ -74,28 +74,13 @@ FAKE_WARP="${TMP_DIR}/fake-warp"
cat >"$FAKE_WARP" <<'EOF'
#!/usr/bin/env bash
set -euo pipefail
[[ " $* " == *" --analyze.v "* ]]
[[ " $* " == *" --no-color "* ]]
if [[ "${FAKE_WARP_ZERO_LATENCY:-0}" == "1" ]]; then
cat <<'LOG'
Operation: GET. Concurrency: 8. Ran: 7s
Requests considered: 1000:
* Average: 160.00 MiB/s, 40960.00 obj/s
* Avg: 0s, 50%: 0s, 90%: 0s, 99%: 0s, Fastest: 0s, Slowest: 1ms, StdDev: 0s
LOG
exit 0
fi
cat <<'LOG'
- PUT Average: 161 Obj/s, 5.0MiB/s; Current 161 Obj/s, 5.0MiB/s.
Operation: GET. Concurrency: 64. Ran: 7s
Requests considered: 1000:
Report: GET. Concurrency: 64. Ran: 7s
* Average: 653.90 MiB/s, 20925.58 obj/s
* Avg: 3.5ms, 50%: 2.0ms, 90%: 3.6ms, 99%: 24.1ms, Fastest: 0.2ms, Slowest: 607.7ms, StdDev: 20.6ms
* Reqs: Avg: 3.5ms, 50%: 2.0ms, 90%: 3.6ms, 99%: 24.1ms, Fastest: 0.2ms, Slowest: 607.7ms, StdDev: 20.6ms
Throughput, split into 7 x 1s:
LOG
if [[ "${FAKE_WARP_ERRORS:-0}" == "1" ]]; then
echo 'Total Errors: 1.'
fi
EOF
chmod +x "$FAKE_WARP"
@@ -118,32 +103,6 @@ chmod +x "$FAKE_WARP"
rg -q '^32767B,warp,1,1,128,ok,0,[^,]+,[^,]+,653.90 MiB/s,685663846.400000,20925.58,3.5 ms,3.500000,[^,]+,3.6 ms,3.600000,24.1 ms,24.100000$' "${TMP_DIR}/fake-warp-run/round_results.csv"
cat >"${TMP_DIR}/warp-no-details.log" <<'EOF'
warp: Starting benchmark in 3s...
Operation: PUT. Concurrency: 8
* Average: 2.76 MiB/s, 707.03 obj/s
EOF
"$RUNNER" --extract-metrics-from-log "${TMP_DIR}/warp-no-details.log" >"${TMP_DIR}/warp-no-details.csv"
rg -qx '2.76 MiB/s,2894069.760000,707.03,N/A,N/A,N/A,N/A,N/A,N/A' "${TMP_DIR}/warp-no-details.csv"
if FAKE_WARP_ERRORS=1 "$RUNNER" \
--tool warp \
--endpoint http://127.0.0.1:9000 \
--access-key test-access \
--secret-key test-secret \
--sizes 32767B \
--rounds 1 \
--retry-per-round 1 \
--retry-sleep-secs 1 \
--cooldown-secs 0 \
--duration 1s \
--out-dir "${TMP_DIR}/fake-warp-errors" \
--warp-bin "$FAKE_WARP" >/dev/null 2>&1; then
echo "expected Warp request errors to fail the benchmark" >&2
exit 1
fi
rg -q ',failed,1,' "${TMP_DIR}/fake-warp-errors/round_results.csv"
"$RUNNER" \
--tool warp \
--endpoint http://127.0.0.1:9000 \
@@ -173,28 +132,4 @@ awk -F',' '
END { exit found ? 0 : 1 }
' "${TMP_DIR}/fake-warp-candidate/baseline_compare.csv"
for leg in baseline candidate; do
zero_args=(
--tool warp
--endpoint http://127.0.0.1:9000
--access-key test-access
--secret-key test-secret
--sizes 4KiB
--rounds 1
--retry-per-round 1
--cooldown-secs 0
--duration 1s
--out-dir "${TMP_DIR}/fake-warp-zero-${leg}"
--warp-bin "$FAKE_WARP"
)
if [[ "$leg" == "candidate" ]]; then
zero_args+=(--baseline-csv "${TMP_DIR}/fake-warp-zero-baseline/median_summary.csv")
fi
FAKE_WARP_ZERO_LATENCY=1 "$RUNNER" "${zero_args[@]}" >/dev/null 2>&1
done
"${SCRIPT_DIR}/hotpath_warp_ab_gate.sh" \
--compare-csv "${TMP_DIR}/fake-warp-zero-candidate/baseline_compare.csv" \
--require-tail-error >/dev/null
echo "object batch benchmark enhanced tests passed"