disttask: refine scheduler error handling #47313

ywqzzy · 2023-09-27T03:59:19Z

What problem does this PR solve?

Issue Number: ref #46258

Problem Summary:
When tidb gracefully shutdown, dist task failed.
When network partition with tikv, dist task failed due to retryable error.

What is changed and how it works?

use backoff for functions related to tikv in scheduler.
check if error retryable, if retryable, don't mark subtask as failed.
use cancelCauseFunc to cancel subtasks. Then when tidb gracefully shutdown, subtask will not be marked as canceled.

func (s *BaseScheduler) markTaskCancelOrFailed(ctx context.Context, subtask *proto.Subtask) bool {
	if err := s.getError(); err != nil {
		if ctx.Err() != nil && context.Cause(ctx).Error() == "cancel subtasks" {
			logutil.Logger(s.logCtx).Warn("subtask canceled", zap.Error(err))
			s.updateSubtaskStateAndError(subtask, proto.TaskStateCanceled, nil)
		} else if common.IsRetryableError(err) {
			logutil.Logger(s.logCtx).Warn("met retryable error", zap.Error(err))
		} else if errors.Cause(err) != context.Canceled {
			logutil.Logger(s.logCtx).Warn("subtask failed", zap.Error(err))
			s.updateSubtaskStateAndError(subtask, proto.TaskStateFailed, err)
		} else {
			logutil.Logger(s.logCtx).Warn("met context canceled for gracefully shutdown", zap.Error(err))
		}
		s.markErrorHandled()
		return true
	}
	return false
}

We use the above function to persist error to subtasks.

Check List

Tests

Unit test
Integration test
Manual test (add detailed scripts or steps below)
No need to test
- I checked and no code files have been changed.

HA test tidb to all network partition passed. Side effects

Performance regression: Consumes more CPU
Performance regression: Consumes more Memory
Breaking backward compatibility

Documentation

Release note

Please refer to Release Notes Language Style Guide to write a quality release note.

None

…koffer

ti-chi-bot · 2023-09-27T03:59:29Z

Hi @ywqzzy. Thanks for your PR.

I'm waiting for a pingcap member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

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/test-infra repository.

tiprow · 2023-09-27T03:59:38Z

Hi @ywqzzy. Thanks for your PR.

PRs from untrusted users cannot be marked as trusted with /ok-to-test in this repo meaning untrusted PR authors can never trigger tests themselves. Collaborators can still trigger tests on the PR using /test all.

I understand the commands that are listed here.

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/test-infra repository.

ywqzzy · 2023-09-27T04:07:31Z

/label ok-to-test

ywqzzy · 2023-09-27T04:07:37Z

/test all

codecov · 2023-09-27T04:27:14Z

Codecov Report

Merging #47313 (4d5501a) into master (0fd232f) will increase coverage by 0.7698%.
The diff coverage is 82.3008%.

Additional details and impacted files

@@               Coverage Diff                @@
##             master     #47313        +/-   ##
================================================
+ Coverage   71.9864%   72.7562%   +0.7698%     
================================================
  Files          1353       1375        +22     
  Lines        401002     407418      +6416     
================================================
+ Hits         288667     296422      +7755     
+ Misses        92965      92204       -761     
+ Partials      19370      18792       -578

Flag	Coverage Δ
integration	`40.1958% <0.8849%> (?)`
unit	`72.0036% <82.3008%> (+0.0172%)`	⬆️

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

Components	Coverage Δ
dumpling	`53.9913% <ø> (ø)`
parser	`84.7472% <ø> (+0.0107%)`	⬆️
br	`49.0100% <ø> (-4.3363%)`	⬇️

…koffer

…se_backoffer

ywqzzy · 2023-10-10T09:52:29Z

/retest

wjhuang2016 · 2023-10-11T06:49:45Z

disttask/framework/scheduler/scheduler.go

@@ -168,6 +186,10 @@ func (s *BaseScheduler) run(ctx context.Context, task *proto.Task) error {
 		proto.TaskStatePending, proto.TaskStateRunning)
 	if err != nil {
 		s.onError(err)


Do we need to record the error if it is retryable?

Do we need to record the error if it is retryable?

No, this will make the dispatcher logic much more complicated, dispatcher need to distinguish the error type from string.
I think we can just check retryable in scheduler.

Benjamin2037 · 2023-10-11T07:51:36Z

disttask/framework/scheduler/manager.go

@@ -67,7 +66,7 @@ type Manager struct {
 		sync.RWMutex
 		// taskID -> cancelFunc.
 		// cancelFunc is used to fast cancel the scheduler.Run.


Does the comment out of date?

Yes, will fix it

Benjamin2037

LGTM

…koffer

D3Hunter

rest lgtm

D3Hunter · 2023-10-12T03:18:58Z

disttask/framework/scheduler/scheduler.go

 	subtask.State = proto.TaskStateSucceed
 	metrics.IncDistTaskSubTaskCnt(subtask)
 }
+
+func (s *BaseScheduler) markTaskCancelOrFailed(ctx context.Context, subtask *proto.Subtask) bool {


Suggested change

func (s *BaseScheduler) markTaskCancelOrFailed(ctx context.Context, subtask *proto.Subtask) bool {

func (s *BaseScheduler) markSubtaskCancelOrFailed(ctx context.Context, subtask *proto.Subtask) bool {

…koffer

D3Hunter · 2023-10-12T04:37:47Z

/lgtm

ti-chi-bot · 2023-10-12T04:37:57Z

[LGTM Timeline notifier]

Timeline:

2023-10-11 08:03:33.247311794 +0000 UTC m=+1212210.834421925: ☑️ agreed by Benjamin2037.
2023-10-12 04:37:54.838636497 +0000 UTC m=+1286272.425746654: ☑️ agreed by D3Hunter.

ti-chi-bot · 2023-10-12T04:37:57Z

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: Benjamin2037, D3Hunter

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

The pull request process is described here

Needs approval from an approver in each of these files:

~~OWNERS~~ [D3Hunter]

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

tiprow · 2023-10-12T04:54:59Z

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

Test name	Commit	Details	Required	Rerun command
tiprow_fast_test	`4d5501a`	link	true	`/test tiprow_fast_test`

Full PR test history. Your PR dashboard.

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/test-infra repository. I understand the commands that are listed here.

ywqzzy added 6 commits September 26, 2023 18:18

fix

b34b940

Merge branch 'master' of https://github.com/pingcap/tidb into use_bac…

74c2dfe

…koffer

fix

ecedd73

Merge branch 'master' of https://github.com/pingcap/tidb into use_bac…

8e01c51

…koffer

fix test

79416cc

fix

7edbced

ti-chi-bot bot added do-not-merge/needs-linked-issue do-not-merge/needs-tests-checked release-note-none Denotes a PR that doesn't merit a release note. needs-ok-to-test Indicates a PR created by contributors and need ORG member send '/ok-to-test' to start testing. labels Sep 27, 2023

ti-chi-bot bot added the size/L Denotes a PR that changes 100-499 lines, ignoring generated files. label Sep 27, 2023

refine

f2648eb

ti-chi-bot bot removed the do-not-merge/needs-tests-checked label Sep 27, 2023

ti-chi-bot bot added the ok-to-test Indicates a PR is ready to be tested. label Sep 27, 2023

ti-chi-bot bot removed the needs-ok-to-test Indicates a PR created by contributors and need ORG member send '/ok-to-test' to start testing. label Sep 27, 2023

ywqzzy added 4 commits September 27, 2023 14:44

fix

224c1db

add cancel cause

be2eb7a

add cancel cause

2c38a4f

retryable error

f64af12

ti-chi-bot bot added size/XL Denotes a PR that changes 500-999 lines, ignoring generated files. and removed size/L Denotes a PR that changes 100-499 lines, ignoring generated files. labels Oct 8, 2023

ywqzzy added 4 commits October 9, 2023 15:17

Merge branch 'master' of https://github.com/pingcap/tidb into use_bac…

6d89e08

…koffer

log refine

17a72b9

fix gracefully shutdown

513b7fd

Merge branch 'master' of https://github.com/pingcap/tidb into use_bac…

e380b17

…koffer

ywqzzy added 6 commits October 10, 2023 16:57

fix

3931c39

Merge branch 'use_backoffer' of https://github.com/ywqzzy/tidb into u…

652e1e7

…se_backoffer

address comments

0968b82

check nil

01550ba

fix spell

e730ac7

rm

e188b34

ywqzzy requested a review from D3Hunter October 10, 2023 09:29

ywqzzy mentioned this pull request Oct 11, 2023

Enhance distributed task execution framework #46258

Closed

76 tasks

wjhuang2016 reviewed Oct 11, 2023

View reviewed changes

Benjamin2037 reviewed Oct 11, 2023

View reviewed changes

Benjamin2037 approved these changes Oct 11, 2023

View reviewed changes

ti-chi-bot bot added the needs-1-more-lgtm Indicates a PR needs 1 more LGTM. label Oct 11, 2023

ywqzzy requested a review from wjhuang2016 October 11, 2023 09:26

ywqzzy added 2 commits October 11, 2023 17:27

Merge branch 'master' of https://github.com/pingcap/tidb into use_bac…

1a6783a

…koffer

fix comments

8f2d96a

D3Hunter reviewed Oct 12, 2023

View reviewed changes

ywqzzy added 3 commits October 12, 2023 12:31

fix

befaba5

add comments

6ee2cac

Merge branch 'master' of https://github.com/pingcap/tidb into use_bac…

4d5501a

…koffer

ti-chi-bot bot added lgtm and removed needs-1-more-lgtm Indicates a PR needs 1 more LGTM. labels Oct 12, 2023

ti-chi-bot bot added the approved label Oct 12, 2023

ti-chi-bot bot merged commit 0c34940 into pingcap:master Oct 12, 2023
12 of 16 checks passed

This was referenced Nov 15, 2023

add index task failed with “Error 1105 (HY000): context canceled” #46777

Closed

add index job rollback after inject one of tidb failure for 1mins or tidb rolling restart #47122

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

disttask: refine scheduler error handling #47313

disttask: refine scheduler error handling #47313

ywqzzy commented Sep 27, 2023 •

edited

Loading

ti-chi-bot bot commented Sep 27, 2023

tiprow bot commented Sep 27, 2023

ywqzzy commented Sep 27, 2023

ywqzzy commented Sep 27, 2023

codecov bot commented Sep 27, 2023 •

edited

Loading

ywqzzy commented Oct 10, 2023

wjhuang2016 Oct 11, 2023

ywqzzy Oct 11, 2023

Benjamin2037 Oct 11, 2023

ywqzzy Oct 11, 2023

Benjamin2037 left a comment

D3Hunter left a comment

D3Hunter Oct 12, 2023

D3Hunter commented Oct 12, 2023

ti-chi-bot bot commented Oct 12, 2023

ti-chi-bot bot commented Oct 12, 2023

tiprow bot commented Oct 12, 2023

	func (s BaseScheduler) markTaskCancelOrFailed(ctx context.Context, subtask proto.Subtask) bool {
	func (s BaseScheduler) markSubtaskCancelOrFailed(ctx context.Context, subtask proto.Subtask) bool {

disttask: refine scheduler error handling #47313

disttask: refine scheduler error handling #47313

Conversation

ywqzzy commented Sep 27, 2023 • edited Loading

What problem does this PR solve?

What is changed and how it works?

Check List

Release note

ti-chi-bot bot commented Sep 27, 2023

tiprow bot commented Sep 27, 2023

ywqzzy commented Sep 27, 2023

ywqzzy commented Sep 27, 2023

codecov bot commented Sep 27, 2023 • edited Loading

Codecov Report

ywqzzy commented Oct 10, 2023

wjhuang2016 Oct 11, 2023

Choose a reason for hiding this comment

ywqzzy Oct 11, 2023

Choose a reason for hiding this comment

Benjamin2037 Oct 11, 2023

Choose a reason for hiding this comment

ywqzzy Oct 11, 2023

Choose a reason for hiding this comment

Benjamin2037 left a comment

Choose a reason for hiding this comment

D3Hunter left a comment

Choose a reason for hiding this comment

D3Hunter Oct 12, 2023

Choose a reason for hiding this comment

D3Hunter commented Oct 12, 2023

ti-chi-bot bot commented Oct 12, 2023

[LGTM Timeline notifier]

ti-chi-bot bot commented Oct 12, 2023

tiprow bot commented Oct 12, 2023

ywqzzy commented Sep 27, 2023 •

edited

Loading

codecov bot commented Sep 27, 2023 •

edited

Loading