[Flaky test] coreinternal/consumerretry/logs_test.go failed on windows

[songy23]

Component(s)

internal/core

What happened?

internal/coreinternal/consumerretry/logs_test.go is broken on Windows test matrix on mainline https://github.com/open-telemetry/opentelemetry-collector-contrib/actions/workflows/build-and-test-windows.yml?query=branch%3Amain

--- FAIL: TestConsumeLogs (0.12s)
    --- FAIL: TestConsumeLogs/retry_success (0.10s)
        logs_test.go:66: 
            	Error Trace:	D:/a/opentelemetry-collector-contrib/opentelemetry-collector-contrib/internal/coreinternal/consumerretry/logs_test.go:66
            	Error:      	Not equal: 
            	            	expected: <nil>(<nil>)
            	            	actual  : *errors.errorString(&errors.errorString{s:"retry later"})
            	Test:       	TestConsumeLogs/retry_success
--- FAIL: TestConsumeLogs_ContextDeadline (0.07s)
    logs_test.go:92: 
        	Error Trace:	D:/a/opentelemetry-collector-contrib/opentelemetry-collector-contrib/internal/coreinternal/consumerretry/logs_test.go:92
        	Error:      	"retry later" does not contain "context is cancelled or timed out retry later"
        	Test:       	TestConsumeLogs_ContextDeadline
FAIL
FAIL	github.com/open-telemetry/opentelemetry-collector-contrib/internal/coreinternal/consumerretry	0.569s

Collector version

mainline

Environment information

Environment

OS: Windows

OpenTelemetry Collector configuration

No response

Log output

No response

Additional context

No response

[github-actions]

Pinging code owners for internal/core: @open-telemetry/collector-approvers. See Adding Labels via Comments if you do not have permissions to add labels yourself.

[github-actions]

Pinging code owners:

• internal/core: @open-telemetry/collector-approvers

See Adding Labels via Comments if you do not have permissions to add labels yourself.

[pjanotti]

I tried multiple runs on my box and got no repro. However, by code inspection I can see, that due to some concurrency on the test code, how both errors could happen. In theory the fix can happen on any system not only Windows. The fix should be entirely on the test code. Giving that this one seems hard to repro I'm giving it a low priority on my work queue, if some volunteer wants to fix them, please do so.

[crobert-1]

Removing needs triage based on @pjanotti's comment.

[pjanotti]

Hit the first one on my fork - documenting here to give a sense of frequency so we can prioritize them accordingly

https://github.com/pjanotti/opentelemetry-service-contrib/actions/runs/6789094249/job/18455541281#:~:text=%2D%2D%2D%20FAIL%3A%20TestConsumeLogs%20(0.12s,588

--- FAIL: TestConsumeLogs (0.12s)
    --- FAIL: TestConsumeLogs/retry_success (0.11s)
        logs_test.go:66: 
            	Error Trace:	D:/a/opentelemetry-service-contrib/opentelemetry-service-contrib/internal/coreinternal/consumerretry/logs_test.go:66
            	Error:      	Not equal: 
            	            	expected: <nil>(<nil>)
            	            	actual  : *errors.errorString(&errors.errorString{s:"retry later"})
            	Test:       	TestConsumeLogs/retry_success
FAIL
FAIL	github.com/open-telemetry/opentelemetry-collector-contrib/internal/coreinternal/consumerretry	0.251s

[pjanotti]

Another hit https://github.com/open-telemetry/opentelemetry-collector-contrib/actions/runs/6797762319/job/18480554554#:~:text=%2D%2D%2D%20FAIL%3A%20TestConsumeLogs%20(0.19s,592

[github-actions]

This issue has been inactive for 60 days. It will be closed in 60 days if there is no activity. To ping code owners by adding a component label, see Adding Labels via Comments, or if you are unsure of which component this issue relates to, please ping @open-telemetry/collector-contrib-triagers. If this issue is still relevant, please ping the code owners or leave a comment explaining why it is still relevant. Otherwise, please close it.

Pinging code owners:

• internal/core: @open-telemetry/collector-approvers

See Adding Labels via Comments if you do not have permissions to add labels yourself.

[pjanotti]

It keeps happening, e.g.: https://github.com/open-telemetry/opentelemetry-collector-contrib/actions/runs/7631311059/job/20788982979#step:7:601

[david-sevcik]

Hello @pjanotti, is it still happening? I would like to take a look at this.

I tried multiple runs on my box and got no repro. However, by code inspection I can see, that due to some concurrency on the test code, how both errors could happen. In theory the fix can happen on any system not only Windows. The fix should be entirely on the test code. Giving that this one seems hard to repro I'm giving it a low priority on my work queue, if some volunteer wants to fix them, please do so.

According to your comment there should be some buggy concurrency code in the test not in the production code right? I don't see any concurrency in the test, can you point me to there please?

Or if I misinterpreted that, I apologize.

[pjanotti]

Hi @david-sevcik - I don't think this one is happening at this time. We can always re-open if appropriate.