You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Linux container in kubernetes in Google Cloud, Eclipse JDK 17
What did you expect to see?
We start up to 20 long-running gRPC requests concurrently, using a fixed thread pool. We use service.withDeadlineAfter(deadline.getSeconds(), TimeUnit.SECONDS) to time out the requests after 15 minutes. If the requests time out we split the work up into smaller requests.
In versions up to and including 1.51.3, this works fine. In versions 1.52.0 and above, the initial requests with the 15 minute deadline never time out.
What did you see instead?
Expected the requests to time out like before.
Steps to reproduce the bug
Please let me know any additional information I can provide.
Looking at the changelogs, it seems maybe this change could be related? #9626
The text was updated successfully, but these errors were encountered:
Yeah, my first thought was #10336. I do know we looked into it more in the past month, but didn't find the direct cause.
I'm going to assume this a duplicate of #10336, as the more we investigate the easier it appears to be hit. If you happen to disable retries and it still happens, then comment here and we can reopen. But I do think it is a safe assumption it is the same issue.
What version of gRPC-Java are you using?
1.51.3, 1.52.0 and 1.57.2
What is your environment?
Linux container in kubernetes in Google Cloud, Eclipse JDK 17
What did you expect to see?
We start up to 20 long-running gRPC requests concurrently, using a fixed thread pool. We use
service.withDeadlineAfter(deadline.getSeconds(), TimeUnit.SECONDS)
to time out the requests after 15 minutes. If the requests time out we split the work up into smaller requests.In versions up to and including 1.51.3, this works fine. In versions 1.52.0 and above, the initial requests with the 15 minute deadline never time out.
What did you see instead?
Expected the requests to time out like before.
Steps to reproduce the bug
Please let me know any additional information I can provide.
Looking at the changelogs, it seems maybe this change could be related? #9626
The text was updated successfully, but these errors were encountered: