gc_state_manager: Make methods of GCStateManager able to timeout or cancel by MyonKeminta · Pull Request #10606 · tikv/pd

MyonKeminta · 2026-04-20T13:48:21Z

What problem does this PR solve?

Issue Number: Close #10607

What is changed and how does it work?

In this PR:

Implemented a simple cancellable RWMutex by using semaphore internally
Added OrderedSingleFlightGroup
In GCStateManager, Replace the original RWLock with the CancellableRWMutex, and set timeout for its APIs
Use read lock instead of write lock in GetGCState

We've encountered a case that some GC APIs are accumulating waiting for the mutex. By making these changes, the case should be able to be alleviated.

gc_state_manager: Make methods of GCStateManager able to timeout or cancel

* Implemented a simple cancellable `RWMutex` by using semaphore internally
* Added `OrderedSingleFlightGroup`
* In `GCStateManager`, Replace the original `RWLock` with the `CancellableRWMutex`, and set timeout for its APIs
* Use read lock instead of write lock in GetGCState

Check List

Tests

Unit test
Integration test
Manual test (add detailed scripts or steps below)
No code

Code changes

Has the configuration change
Has HTTP APIs changed (Don't forget to add the declarative for the new API)
Has persistent data change

Side effects

Possible performance regression
Increased code complexity
Breaking backward compatibility

Related changes

PR to update pingcap/docs/pingcap/docs-cn:
PR to update pingcap/tiup:
Need to cherry-pick to the release branch

Release note

None.

Signed-off-by: MyonKeminta <MyonKeminta@users.noreply.github.com>

ti-chi-bot · 2026-04-20T13:48:25Z

Skipping CI for Draft Pull Request.
If you want CI signal for your change, please convert it to an actual PR.
You can still manually trigger a test run with /test all

ti-chi-bot · 2026-04-20T13:48:28Z

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by:
Once this PR has been reviewed and has the lgtm label, please assign hundundm for approval. For more information see the Code Review Process.
Please ensure that each of them provides their approval before proceeding.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

coderabbitai · 2026-04-20T13:48:30Z

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: a94a52c3-2cc5-4cf9-ac46-8210016ca104

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

Create PR with unit tests

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

MyonKeminta · 2026-04-20T14:39:16Z

/retest

ti-chi-bot · 2026-04-20T14:50:12Z

@MyonKeminta: The following tests failed, say /retest to rerun all failed tests or /retest-required to rerun all mandatory failed tests:

Test name	Commit	Details	Required	Rerun command
pull-unit-test-next-gen-1	`aff7ba3`	link	true	`/test pull-unit-test-next-gen-1`
pull-unit-test-next-gen-3	`aff7ba3`	link	true	`/test pull-unit-test-next-gen-3`
pull-unit-test-next-gen-2	`aff7ba3`	link	true	`/test pull-unit-test-next-gen-2`

Full PR test history. Your PR dashboard.

Details

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. I understand the commands that are listed here.

codecov · 2026-04-20T14:53:22Z

Codecov Report

❌ Patch coverage is 81.39535% with 40 lines in your changes missing coverage. Please review.
✅ Project coverage is 78.98%. Comparing base (b21a183) to head (aff7ba3).
⚠️ Report is 2 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master   #10606      +/-   ##
==========================================
+ Coverage   78.96%   78.98%   +0.01%     
==========================================
  Files         532      533       +1     
  Lines       71883    72123     +240     
==========================================
+ Hits        56766    56965     +199     
- Misses      11093    11114      +21     
- Partials     4024     4044      +20

Flag	Coverage Δ
unittests	`78.98% <81.39%> (+0.01%)`	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

rleungx · 2026-04-20T15:00:01Z

 	// layer. It can be more efficient and avoid failures due to transaction conflict in most cases.
 	// The etcd transactions is still necessary considering the possibility of rare cases like PD leader changes.
-	mu              syncutil.RWMutex
+	mu              syncutil.CancellableRWMutex


Why do we need a cancellable mutex?

As it may involve etcd IO in its critical section, and it may need a long time to successfully acquire the lock. Without the ability to cancel or timeout, the goroutines waiting for it would be unable to exit, and the memory used by them can't be released either.

rleungx · 2026-04-20T15:23:04Z

+	})
+}
+
+func (m *GCStateManager) getGCStateImpl(ctx context.Context, keyspaceID uint32) (GCState, error) {


When the context is cancelled, it cannot cancel the transaction.

This is a problem but it's unable to pass ctx to the StorageEndpoint API now. Seems it requires a large refactor.

rleungx · 2026-04-20T15:24:06Z

+
+// RLock acquires a read lock or returns an error if ctx is canceled before the
+// lock is acquired.
+func (m *CancellableRWMutex) RLock(ctx context.Context) error {


The deadlock testing does not work anymore.

rleungx · 2026-04-20T15:26:37Z

+
+func (m *CancellableRWMutex) init() {
+	m.initOnce.Do(func() {
+		m.inner = semaphore.NewWeighted(maxReaderCount)


Will the read be blocked more easily?

The overhead of the lock itself is larger than the builtin RWMutex, but in the current use case I think it won't

MyonKeminta added 2 commits April 20, 2026 16:55

Add cancellable rw mutex

e67ec4a

Signed-off-by: MyonKeminta <MyonKeminta@users.noreply.github.com>

Add timeout to methods of GCStateManager

aff7ba3

Signed-off-by: MyonKeminta <MyonKeminta@users.noreply.github.com>

ti-chi-bot Bot added do-not-merge/needs-linked-issue do-not-merge/work-in-progress Indicates that a PR should not merge because it is a work in progress. release-note-none Denotes a PR that doesn't merit a release note. labels Apr 20, 2026

ti-chi-bot Bot added the dco-signoff: yes Indicates the PR's author has signed the dco. label Apr 20, 2026

ti-chi-bot Bot added size/XXL Denotes a PR that changes 1000+ lines, ignoring generated files. and removed do-not-merge/needs-linked-issue labels Apr 20, 2026

rleungx reviewed Apr 20, 2026

View reviewed changes

Conversation

MyonKeminta commented Apr 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What problem does this PR solve?

What is changed and how does it work?

Check List

Release note

Uh oh!

ti-chi-bot Bot commented Apr 20, 2026

Uh oh!

ti-chi-bot Bot commented Apr 20, 2026

Uh oh!

coderabbitai Bot commented Apr 20, 2026

Review skipped

Uh oh!

MyonKeminta commented Apr 20, 2026

Uh oh!

ti-chi-bot Bot commented Apr 20, 2026

Uh oh!

codecov Bot commented Apr 20, 2026

Codecov Report

Uh oh!

rleungx Apr 20, 2026

Choose a reason for hiding this comment

Uh oh!

MyonKeminta Apr 21, 2026

Choose a reason for hiding this comment

Uh oh!

rleungx Apr 20, 2026

Choose a reason for hiding this comment

Uh oh!

MyonKeminta Apr 21, 2026

Choose a reason for hiding this comment

Uh oh!

rleungx Apr 20, 2026

Choose a reason for hiding this comment

Uh oh!

rleungx Apr 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

MyonKeminta Apr 21, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

MyonKeminta commented Apr 20, 2026 •

edited

Loading

rleungx Apr 20, 2026 •

edited

Loading