-
Notifications
You must be signed in to change notification settings - Fork 1.3k
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
[prebuild][selfhosted] prebuilds freeze and does not fail nor finish. #8942
Comments
Hey @fredrikhgrelland, is repo public? It would help to reproduce. (maybe one of your dockerfile commands requires user input?) |
@axonasif I made a reproducible repo public here: https://github.com/fredrikhgrelland/gitpod-public-sandbox. Case:
|
Hey @fredrikhgrelland, Thanks for the detailed issue, repro cases. Definitely helps make debugging easier. :) So, I have a
Debugging more on this, and will update on this once I find more! but the prebuild should still run to completion (just like in my case) and I'm wondering why that is not the case in your environment 🤔 |
Yeah, it is bit of a head-scratcher. Are you seeing these kind of logs as well: |
I have run some more tests and I have maybe found something by looking at the prebuild-pod logs in real-time.
While logs are streaming running for ~25 seconds:
Log-streaming freeze and this pops up in the logs:
The prebuild will now hang for ~1 hour and then continue to a image-build stage until starting the workspace, the prebuild clearly not done. NOTE: Normally it gets in a freeze-failing state far earlier and before the web-app starts printing logs. |
I have some more observations that might help further: The webapp log-streaming as well as attaching to the prebuild pod tailing from |
That is a UX issue. I see it from time to time as well: Prebuilds are running but the logs are not shown in the UI. Reloading the window helps for me usually. I think it's issue #8684, correct me when I'm wrong, @geropl.
For prebuilds, we have a timeout of 1 hour. How long does the build usually take? |
Exactly. #8684 does overlap with other issues already merged (but not deployed yet), so: we're working on this part. @fredrikhgrelland Thanks for your very detailed analysis! 🙏 As far as I understand, especially after reading this comment of yours, this is a supervisor/workspace problem:
@kylos101 This sounds like a runtime issue. Can someone from workspace have a look at this? @sagor999 maybe? 🤔 |
Hi @fredrikhgrelland 👋 , Can you tell if you're hitting this line? If yes, may I ask you to confirm from a regular workspace, what version of I ask because about a month ago we updated the version of docker in |
Hi @kylos101,
Unless you are running the
|
Is there any more steps I can take in order to debug a potential supervisor/workspace-issue? |
@fredrikhgrelland What version of containerd are you on? 🤔 |
@geropl I am running |
Hey @fredrikhgrelland , are you still having trouble with prebuilds freezing? CC: @metcalfc |
Hi @fredrikhgrelland , I am going to close this for now, but, if it's still a problem, let us know and we'll happily reopen. 🙏 CC: @metcalfc 👀 |
Bug description
I am running a docker build as an init-task in a self-hosted gitpod on EKS.
This works fine every time when opening a new workspace.
I have given the pods 16gb ram and 4 cpus.
When starting a prebuild with with the magic-link
https://gitpod.example.com/#prebuild/https://github/myrepo
one of the following things will happen:For both cases i can observe:
supervisor init
tail -f /workspace/.gitpod/prebuild-log-1
seems to stop at a random place in the process. (the log below came from a prebuild that never showed any logs in the web-uiSteps to reproduce
Running a prebuild building a docker image with significant load?
Workspace affected
No response
Expected behavior
Prebuilds should fail or succed. Now stuck in limbo.
Example repository
No response
Anything else?
No response
The text was updated successfully, but these errors were encountered: