-
Notifications
You must be signed in to change notification settings - Fork 39.5k
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
TestWatch {etcd} #31262
Comments
This blocks the queue it seems. |
Assigning according to this. |
I haven't been able to reproduce locally, but I do see some odd output. It appears that TestWatch tries to connect to the sockets for the etcd nodes created in the other tests in this file.
|
https://k8s-gubernator.appspot.com/build/kubernetes-jenkins/logs/kubernetes-test-go/16144/ Failed: TestWatch {etcd}
|
https://k8s-gubernator.appspot.com/build/kubernetes-jenkins/logs/kubernetes-test-go/16148/ Failed: TestWatch {etcd}
|
https://k8s-gubernator.appspot.com/build/kubernetes-jenkins/logs/kubernetes-test-go/16156/ Failed: TestWatch {etcd}
|
https://k8s-gubernator.appspot.com/build/kubernetes-jenkins/logs/kubernetes-test-go/16162/ Failed: TestWatch {etcd}
|
https://k8s-gubernator.appspot.com/build/kubernetes-jenkins/logs/kubernetes-test-go/16164/ Failed: TestWatch {etcd}
|
This is definitely related to this PR: that switched unit tests to use v3 API of etcd. @hongchaodeng @xiang90 @timothysc - seems like a blocker. @lavalamp - FYI |
https://k8s-gubernator.appspot.com/build/kubernetes-jenkins/logs/kubernetes-test-go/16169/ Failed: TestWatch {etcd}
|
FWIW I ran unit-tests locally tens of times without failure. I wonder if it's similar to the resource issue we found a while ago on our jenkins env /cc @ixdy |
https://k8s-gubernator.appspot.com/build/kubernetes-jenkins/logs/kubernetes-test-go/16183/ Failed: TestWatch {etcd}
|
I made a rollback PR. #31371 |
Some findings on logs from last failure -- the test infra is leaking clients Here are the logs for failed test:
Basically, |
After digging into the code, we are obviously not closing any clients in a lot of tests. |
Submit a fix: #31390 |
https://k8s-gubernator.appspot.com/build/kubernetes-jenkins/logs/kubernetes-test-go/16194/ Failed: TestWatch {etcd}
|
https://k8s-gubernator.appspot.com/build/kubernetes-jenkins/logs/kubernetes-test-go/16195/ Failed: TestWatch {etcd}
|
https://k8s-gubernator.appspot.com/build/kubernetes-jenkins/logs/kubernetes-test-go/16198/ Failed: TestWatch {etcd}
|
Looks like we stopped seeing these after #31371 merged. |
Automatic merge from submit-queue return destroy func to clean up internal resources of storage What? Provide a destroy func to clean up internal resources of storage. It changes **unit tests** to clean up resources. (Maybe fix integration test in another PR.) Why? Although apiserver is designed to be long running, there are some cases that it's not. See #31262 (comment) We need to gracefully shutdown and clean up resources.
Which was expected BTW :) |
I've nearly got the deferFunc() plumbed through to the rest of the registry testing and will put up a PR, but I'm happy to wait till 1.5. |
fix posted. |
Automatic merge from submit-queue Etcd3 v3client + grpc client leak fix This re-enables the client and plumbs through the destroyFunc to cleanup clients. Fix for #31262
Automatic merge from submit-queue return destroy func to clean up internal resources of storage What? Provide a destroy func to clean up internal resources of storage. It changes **unit tests** to clean up resources. (Maybe fix integration test in another PR.) Why? Although apiserver is designed to be long running, there are some cases that it's not. See kubernetes/kubernetes#31262 (comment) We need to gracefully shutdown and clean up resources.
https://k8s-gubernator.appspot.com/build/kubernetes-jenkins/logs/kubernetes-test-go/16135/
Failed: TestWatch {etcd}
The text was updated successfully, but these errors were encountered: