disktask: replace failure nodes with alive ones

[xhebox]

What problem does this PR solve?

Issue Number: ref #46258

Problem Summary: This PR adds a simple mechanism to cancel existed subtask.

What is changed and how it works?

Dispatcher:

1. fetches liveNodes from etcd every DefaultLiveNodesCheckTick ticks in handleRunning.
2. lazily fetched taskNodes in handleRunning as a local cache.
3. If some of taskNodes are absent from liveNodes, it is marked as failure. And random nodes liveNodes will replace them. It is done by UpdateFailedSchedulerIDs. The local cache is also updated.

Scheduler:

1. starts a background task every Run(..) for checking if the scheduler has been canceled by dispatcher. If so, call runtimeCancel.

Check List

Tests

• Unit test
  • 3 subtasks
    • 4 nodes. 3 nodes are hung up, then the task is forwarded to the rest nodes and executed successfully.
    • 30 nodes. 3 nodes are hung up, then the task is forwarded to the rest nodes and executed successfully.
  • 10 subtasks in the first stage, 5 subtasks in the second stage
    • 4 nodes. 3 nodes are hung up in the first stage, then the task is forwarded to the rest nodes and executed successfully.
    • 30 nodes, 3 nodes are hung up in the first stage, then the task is forwarded to the rest nodes and executed successfully.
    • 6 nodes. 3 nodes are hung up in the first stage, 1 tidb node in the second stage hangs up. The task is forwarded to the rest nodes and executed successfully.
    • 30 nodes, 3 nodes are hung up in the first stage, 1 tidb node in the second stage hangs up. The task is forwarded to the rest nodes and executed successfully.
  • Node network isolated and restored
    • 4 nodes. 3 nodes are isolated from the network and recovered. These nodes no longer pick tasks, and the other nodes execute all subtasks.
    • 30 nodes. 3 nodes are isolated from the network and recovered. These nodes no longer pick tasks, and the other nodes execute all subtasks.
• Integration test
• Manual test (add detailed scripts or steps below)
• No code

Side effects

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

Documentation

• Affects user behaviors
• Contains syntax changes
• Contains variable changes
• Contains experimental features
• Changes MySQL compatibility

Release note

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

None

[ti-chi-bot]

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

[tiprow]

Hi @xhebox. 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.

[xhebox]

/run-all-tests

[codecov]

Codecov Report

Merging #45935 (92ff681) into master (a427113) will decrease coverage by 0.6077%.
Report is 4 commits behind head on master.
The diff coverage is 79.6116%.

Additional details and impacted files

@@               Coverage Diff                @@
##             master     #45935        +/-   ##
================================================
- Coverage   73.3303%   72.7226%   -0.6078%     
================================================
  Files          1310       1331        +21     
  Lines        394940     401568      +6628     
================================================
+ Hits         289611     292031      +2420     
- Misses        86881      91048      +4167     
- Partials      18448      18489        +41     

Flag	Coverage Δ
integration	25.4447% <0.0000%> (?)
unit	73.4115% <83.6734%> (+0.0811%)	⬆️

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

Components	Coverage Δ
dumpling	54.0444% <ø> (ø)
parser	84.9628% <ø> (+0.0108%)	⬆️
br	48.0377% <ø> (-4.2826%)	⬇️

[ywqzzy]

/retest

[tiprow]

@ywqzzy: Cannot trigger testing until a trusted user reviews the PR and leaves an /ok-to-test message.

In response to this:

/retest

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]

/cc @GMHDBJD

[ywqzzy]

/retest

[tiprow]

@ywqzzy: Cannot trigger testing until a trusted user reviews the PR and leaves an /ok-to-test message.

In response to this:

/retest

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.

[ti-chi-bot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: tangenta, ywqzzy

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 [tangenta]

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

[ti-chi-bot]

[LGTM Timeline notifier]

Timeline:

• 2023-08-30 02:55:45.047838637 +0000 UTC m=+1895709.596854624: ☑️ agreed by ywqzzy.
• 2023-09-04 07:01:10.787664033 +0000 UTC m=+2342435.336680021: ☑️ agreed by tangenta.