-
Notifications
You must be signed in to change notification settings - Fork 10.3k
[WIP] Run helix tests on master #6728
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
Yay arcade fix unblocked running helix the real way again, I had to remove the mvc functional tests and the signalr client ts functional tests though since they failed with build errors as part of the helix gather, not sure what's special about those two, I skipped the from building helix payloads for now |
2f3ee15
to
e8f2c85
Compare
src/Components/Blazor/Build/test/RuntimeDependenciesResolverTest.cs
Outdated
Show resolved
Hide resolved
src/Hosting/test/FunctionalTests/Microsoft.AspNetCore.Hosting.FunctionalTests.csproj
Show resolved
Hide resolved
89e7c8c
to
ab7ae5f
Compare
Need any help from SignalR folks to figure this out? |
Sure @BrennanConroy that'd be great, you just need to turn BuildHelixPayload back on for the test project in this branch to investigate |
ff07e13
to
2e4316b
Compare
Helix tests are finally green including localdb tests as well. Functional tests and tests that rely on the sln are the biggest category of SkipOnHelix tests left |
17 test projects are skipped completely on helix (14 functional + 2 mvc analyzers + dotnet watch) + 13 individually skipped tests (mostly looking for sln + cert tests on OSX). Rest of the tests look good now. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It looks like this is really close. I've left a few nit-picks and questions. The CI is failing normal builds with this change (obviously something to fix first), but other than that, I'd say this is ready to go.
...gnalR/server/Specification.Tests/src/Microsoft.AspNetCore.SignalR.Specification.Tests.csproj
Outdated
Show resolved
Hide resolved
src/Identity/Specification.Tests/src/Microsoft.AspNetCore.Identity.Specification.Tests.csproj
Outdated
Show resolved
Hide resolved
@MattGal do you know what I need to do to get helix test failures to get reported to the AzDo UI that nate is mentioning? |
@natemcmaster should we move the helix job to a separate checkbox similar to how pr-fastvalidation used to be instead of running it as part of normal ci to start? |
Yeah, that's probably a good idea while we stabilize Helix. I expect we'll find more flaky tests an account of how many times each test will run. If you move the YML changes to .azure/pipelines/helix-test.yml, I can setup a new AzDO definition. |
@natemcmaster I believe to get this you should only need to set |
@natemcmaster moved the helix job to its own yml file, and also turned on the azure reporting flag. Running time is also approx 20 min per queue, so running all 10 queues takes a while. Matt identified some obvious perf bottlenecks a few weeks ago, I'll see if I can do some of the low hanging ones to speed things up now that things are working |
Thanks @MattGal! @HaoK this looks awesome! Obviously there is a handful of work necessary to get all of our tests running in Helix, but I think we should get this merged soon. I'm about to merge #7137 which may have a few conflicts on ci.yml, but otherwise I think this should be good to merge today. |
Yeah I'll rebase against master again now, I'm not sure why the normal ci would be failing after I do that since the helix stuff shouldn't be affecting it at all anymore |
0dd670e
to
ec6b459
Compare
My changes are in master now. Looks like the merge conflict is on FunctionalTests.csproj now. |
Almost there, helix tests are all passing after skipping a few new failing tests, @natemcmaster can you add the system access token to the pipeline that helix is trying to access? Hopefully things will finally go green again once it has that.
|
Green! 👍 |
Yeah, guess the azure error reporting thing is causing that new failure, I turned that flag off for now. retriggering the normal ci since that failed now :( |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Awesome, let's do this. Let's open a bug to follow up on re-enabling AzDO reporting.
Last suggestions are nits about code organization, but I don't want to block the PR on this since this has been in progress for so long.
Nit 1 - I'd like to avoid adding a new top-level folder to this repo. We are using the eng/
folder for targets/tasks/scripts etc. eng/helix/
or eng/scripts/
seems like a better place for some of this code.
Nit 2 - put shared source code (SkipOnHelixAttribute.cs) in src/Shared/
Cool, I'll merge and start another PR with the tweaks, and it looks like there's a few flaky tests still too since the helix job still fails like half the time. |
Now that this is merged and running on peoples PR's, an email to the team describing what it is, how it works, and how to see failures would be nice. |
Good idea I'll send something out this afternoon |
Awesome to see this merged you guys, you came a long way in a (relatively) short period of time on this. |
17 Skipped projects:
13 Individually Skipped tests