-
Notifications
You must be signed in to change notification settings - Fork 387
Integration tests appear to hold tons of file descriptors open #1649
Comments
I see this too. |
It looks like the culprit is etcd client connections. Seems like etcd is not being shut down after each test case. |
That's neat. Your |
when I run when I run
|
We do not need that anymore with the builtin etcd. |
The storage info object gets sent to genericapiserver.go, InstallAPIGroup. We need to look into if there is a bug where this class never calls the DestroyFunc for storage. |
Also, I was mistaken, we use the same etcd server for all the integration tests. So what we have no is the same as we had before, the main difference is that you don't need to have etcd installed locally to run the unit tests. Note, the embedded etcd will cluster with a local etcd if given a chance. But that is not the issue we are seeing. See these: |
I want a server that cleanly shuts down. We should document what we find. goroutines definietly going up per test case, youch.
|
Digging about more, I also see |
This is still an issue and wasn't meant to be closed |
lol, the logic behind the scenes there. the word "fix" and then a link to the issue. |
If kubernetes/apiserver#30 gets implemented we'd have to update/backport to it to get the functionality. I think with the existing 1.9 release we might be able to do this etcd-watcher cleanup by implementing the destroy on the storage and using a |
more data for runs and incoming pr https://gist.github.com/MHBauer/8c7c94c43f4a53b20ee447a926641166 |
partial mitigation for #1649 still many outstanding goroutines by test end
still threads hanging about, even if I pause for 5minutes after all tests have run. maybe a longer timeout is needed, but this is what I see:
|
Issues go stale after 90d of inactivity. If this issue is safe to close now please do so with Send feedback to sig-testing, kubernetes/test-infra and/or fejta. |
Stale issues rot after 30d of inactivity. If this issue is safe to close now please do so with Send feedback to sig-testing, kubernetes/test-infra and/or fejta. |
Rotten issues close after 30d of inactivity. Send feedback to sig-testing, kubernetes/test-infra and/or fejta. |
@fejta-bot: Closing this issue. In response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
...eventually causing the process to crash on my personal machine.
Output:
The text was updated successfully, but these errors were encountered: