Flakes: Address TestServerStats flakiness

[mattlord]

Description

We have seen the TestServerStats unit test fail pretty regularly, in particular in the unit_race workflow. Example failure:

--- FAIL: TestServerStats (0.02s)
    server_test.go:265: listening on address '127.0.0.1' port 41297
    server_test.go:1363: Running mysql command: /usr/bin/mysql [-v -v -v -e error -h 127.0.0.1 -P 41297 -u user1 -ppassword1]
    server_test.go:670: 
        	Error Trace:	/opt/actions-runner/_work/vitess-private/vitess-private/go/mysql/server_test.go:670
        	Error:      	Not equal: 
        	            	expected: int(0)
        	            	actual  : int64(1)
        	Test:       	TestServerStats
        	Messages:   	connCount

I could not repeat the failure locally (despite trying many different methods), but in walking through the code it was clear that we had a potential race here. The race is this:

• The test starts a mysql server component (to handle the mysql protocol and incoming client connections) and uses mysql cli invocations to connect and execute queries against this internal server
• The server component accepts this incoming connection in Accept
  • And it starts a goroutine to handle this connection, after it's been accepted, in the handle function:

    vitess/go/mysql/server.go

    Lines 327 to 359 in dc692fa

    	// Accept runs an accept loop until the listener is closed.
    	func (l *Listener) Accept() {
    	ctx := context.Background()
    	for {
    	conn, err := l.listener.Accept()
    	if err != nil {
    	// Close() was probably called.
    	connRefuse.Add(1)
    	return
    	}
    	acceptTime := time.Now()
    	connectionID := l.connectionID
    	l.connectionID++
    	connCount.Add(1)
    	connAccept.Add(1)
    	go func() {
    	if l.PreHandleFunc != nil {
    	conn, err = l.PreHandleFunc(ctx, conn, connectionID)
    	if err != nil {
    	log.Errorf("mysql_server pre hook: %s", err)
    	return
    	}
    	}
    	l.handle(conn, connectionID, acceptTime)
    	}()
    	}
    	}

• The test immediately checked some stats that are changed in that handle function running in the goroutine, most notably connCount which is decremented in a defer:

  vitess/go/mysql/server.go

  Lines 390 to 391 in dc692fa

  	// Adjust the count of open connections
  	defer connCount.Add(-1)

So in this PR we wait for the expected value to eliminate this race.

Related Issue(s)

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

[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

All modified and coverable lines are covered by tests ✅

Project coverage is 67.02%. Comparing base (da49d08) to head (364af21).
Report is 5 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #16991      +/-   ##
==========================================
- Coverage   69.35%   67.02%   -2.33%     
==========================================
  Files        1571     1571              
  Lines      204183   251677   +47494     
==========================================
+ Hits       141601   168686   +27085     
- Misses      62582    82991   +20409     

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

[GuptaManan100]

TIL assert.Eventually exists! I was up until now always writing a for loop with a Ticker 😿!