VDiff: fix race when a vdiff resumes on vttablet restart

[rohit-nayak-ps]

Description

There is a race when vttablet starts and the vdiff engine opens. It starts vdiffs which need to be resumed where we restart vdiffs in error state concurrently with the vdiff engine init. It is possible that globalStats map is not initialized when a new controller gets added and the controller adds it self to the map.

This PR starts the goroutine that runs the controller thread only after the inits have been done.

if err := vde.initControllers(rows); err != nil {
		return err
	}
	vde.updateStats()

	// At this point we've fully and successfully opened so begin
	// retrying error'd VDiffs until the engine is closed.
	vde.wg.Add(1)
	go func() {
		defer vde.wg.Done()
		if vde.fortests {
			return
		}
		vde.retryErroredVDiffs()
	}()

This was added way back, so we backport to all supported versions: https://github.com/vitessio/vitess/pull/10639/files#diff-931d120fc70d7007ffd77176f54c526f678c1a6c4e703dbf54b3086c967dbd3dR323

Related Issue(s)

• Fixes: Bug Report: Segfault and panic when restarting tablet with active vdiff #17637

Checklist

• "Backport to:" labels have been added if this change should be back-ported to release branches
• If this change is to be back-ported to previous releases, a justification is included in the PR description
• Tests were added or are not required
• Did the new or modified tests pass consistently locally and on CI?
• Documentation was added or is not required

Deployment Notes

[vitess-bot]

Review Checklist

Hello reviewers! 👋 Please follow this checklist when reviewing this Pull Request.

General

• Ensure that the Pull Request has a descriptive title.
• Ensure there is a link to an issue (except for internal cleanup and flaky test fixes), new features should have an RFC that documents use cases and test cases.

Tests

• Bug fixes should have at least one unit or end-to-end test, enhancement and new features should have a sufficient number of tests.

Documentation

• Apply the release notes (needs details) label if users need to know about this change.
• New features should be documented.
• There should be some code comments as to why things are implemented the way they are.
• There should be a comment at the top of each new or modified test to explain what the test does.

New flags

• Is this flag really necessary?
• Flag names must be clear and intuitive, use dashes (-), and have a clear help text.

If a workflow is added or modified:

• Each item in Jobs should be named in order to mark it as required.
• If the workflow needs to be marked as required, the maintainer team must be notified.

Backward compatibility

• Protobuf changes should be wire-compatible.
• Changes to _vt tables and RPCs need to be backward compatible.
• RPC changes should be compatible with vitess-operator
• If a flag is removed, then it should also be removed from vitess-operator and arewefastyet, if used there.
• vtctl command output order should be stable and awk-able.

[codecov]

Codecov Report

Attention: Patch coverage is 87.50000% with 2 lines in your changes missing coverage. Please review.

Project coverage is 67.73%. Comparing base (9c6c380) to head (853bb6f).

Files with missing lines	Patch %	Lines
go/vt/vttablet/tabletmanager/shard_sync.go	77.77%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #17638      +/-   ##
==========================================
- Coverage   67.76%   67.73%   -0.03%     
==========================================
  Files        1586     1586              
  Lines      255763   255758       -5     
==========================================
- Hits       173315   173246      -69     
- Misses      82448    82512      +64     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[rohit-nayak-ps]

Added this because we also saw a panic here when the vdiff panic occurred. Possibly a different root cause (race?) is causing that since we should never get a nil event, but adding a guard here so vttablet doesn't panic in such a case but logs it.

[mattlord]

Thanks, @rohit-nayak-ps !

[mattlord]

@rohit-nayak-ps the stats were only added in v20: fa59f9d

[mattlord]

This comment is no longer applicable. And we could revert the related changes too. Fine either way IMO.

[mattlord]

The vdiff engine's controller list is initialized as empty in Open and I think that we should do the same for the stats so that they stay in sync and the stats don't grow and grow with old cruft. We can do that by placing this block just after the if len(vde.controllers) > 0 { block above:

	func() {
		globalStats.mu.Lock()
		defer globalStats.mu.Unlock()
		globalStats.controllers = make(map[int64]*controller)
	}()