-
Notifications
You must be signed in to change notification settings - Fork 2.9k
Move our tests to a recent Keycloak version #49283
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
base: main
Are you sure you want to change the base?
Conversation
@sberyozkin @michalvavrik for your awareness, I will need some help at some point but I'll ping you when I have a better idea of the status. |
Sure, thanks for looking at it. FYI - RHSSO has another 2 years in EUS, so I don't know what @sberyozkin wants. I definitely think we should migrate because in product we only support RHBK anyway, we don't test RHSSO nor document it in supported configurations. Anyway, @sberyozkin is the one who must decide. |
Well, tbh, I didn't start this work in a hurry for fun: the Keycloak legacy image is very old and plagued with a JDK bug that is highly problematic for us - the Docker image can't be started on RunsOn anymore and this has been highly problematic for CI. While @crohr from RunsOn might be able to kindly provide us with a temporary workaround, we need to move on. We don't have any recent image to test things with. And we are not actually testing things with recent Keycloak versions, which is IMHO a mistake. If we want to keep supporting old versions with extended support, I think it should be done downstream and not with the community bits where we don't have access to recent images. |
This comment has been minimized.
This comment has been minimized.
🎊 PR Preview a04eb95 has been successfully built and deployed to https://quarkus-pr-main-49283-preview.surge.sh/version/main/guides/
|
@gsmet , ok, I think it is for better to migrate, I just felt obliged to warn you because I did proposed dropping legacy in past and Sergey wished to wait. I expect that your arguments are strong enough. |
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
I didn't run |
@gsmet it's passing for me locally in native, maybe it would be worth to rerun just |
This comment has been minimized.
This comment has been minimized.
I actually can't see how |
I think we would need to print out KC logs to decide what is happening in there, so far my only idea is resources. |
It looks like it tries to connect to Keycloak on an incorrect port. |
I think this is probably right assessment, but I can't tell from logs. There isn't enough information, neither I can say how is it related to changes in this PR (I checked Develocity and it started failing now so probably it is, but I don't know how), |
I have tried it again, this time I switched from Podman to Docker and still can't to reproduce, nor I can see how this failure is related to changes in this PR. I can see you are trying workaround #49291. I don't know how to debug it without using PR CI to debug, so my preferred way would be to disable the test and try to open a new PR right after this is merged to test enabling the test in question. I can't tell how else to debug it. We won't move forward without printing more information. |
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
I'll go back on this again today, if I still can't reproduce it locally, I'll just open PR for CI experiments, because current logs makes me no smarter, the KC port looks fine. |
We are running a very old legacy version for testing which is a bad idea. Especially since this version is using a very old JDK. This is a first pass, it is still very rough and some tests are not passing yet but I would like some feedback from a full CI run. Co-authored-by: Michal Vavřík <mvavrik@redhat.com>
I squashed and made you a coauthor. Let's see how it goes :). |
Status for workflow
|
Status for workflow
|
@sberyozkin so should we get this in? |
@sberyozkin friendly ping? |
I think it will have to wait another week or more considering the situation. |
We are running a very old legacy version for testing which is a bad idea.
Especially since this version is using a very old JDK, which unfortunately is plagued by a bug that prevents it to start on recent Ubuntu runners (https://bugs.openjdk.org/browse/JDK-8287073, thanks to @Ladicek for tracking it).
I have no problem with us having to test this version with our stuff, but I don't think it should be the default in the community and we probably need QE testing on specific machines that will be able to run this old JDK.