-
Notifications
You must be signed in to change notification settings - Fork 7.9k
Stabilizing the FipsDistTest #26401
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
Stabilizing the FipsDistTest #26401
Conversation
ahus1
left a comment
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.
I'd like to see a comment in the code how the 3 minutes are determined.
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.
This is IMHO missing a comment why you whose 3 minutes here. Otherwise people will wonder in the future why this is here.
Looking at application.properties, the transaction timeout is 300 seconds, which is 5 minutes.
Waiting for 3 minutes here, and (given your comment in the issue) waiting for 2 minutes at another place, would add up to 5 minutes, which matches the 300 seconds. Maybe it would be good to add one more minute to he on the safe side. Whatever your reasoning is here, state the "why" for those minutes you choose here in a comment.
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.
Here I wrongly increased the timeout on the kill command while we actually need to increase the timeout for the stop command. The reason is the mentioned Narayana bug, which makes the process hang longer. I already made the change, I also added a comment that the timeout can be reverted once the bug is addressed.
Speaking about the timeout length, 180 seconds should be just enough. From the perspective of other tests, the used .waitFor() will continue as soon as the process terminates. This way the other tests shouldn't be influenced.
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.
@Pepo48 - I see that the code has changed, but I'd still like to see a comment in the code detailing the "why" you chose 180 seconds so our future selfs and other will know.
I someone wants to comment why such a comment is not necessary, I'm open to hear that.
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.
This is in any case just a workaround that should go away once jbosstm/narayana#2201 is resolved.
Couldn't find an issue to remove the workaround. @shawkins Did we create one, I'm not sure. :)
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.
Couldn't find an issue to remove the workaround. @shawkins Did we create one, I'm not sure. :)
Not yet, just the comments in the code about the upstream issue.
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.
Created #26454
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.
@ahus1 As far as I understand, the 180 seconds is just an "educated guess" (as with anything when dealing with async stuff in testsuite). Do you want to add a comment with that? :)
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.
I was hoping something different than an educated guess, let's leave it as is.
* increased the timeout to let Keycloak stop Closes keycloak#26374 Signed-off-by: Peter Zaoral <[email protected]>
shawkins
left a comment
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.
Looks good to me. If test runtimes become an issue we can put in more proactive destroyForcibly calls.
ahus1
left a comment
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.
Going "neutral" after previous discussion.
Closes #26374