-
Notifications
You must be signed in to change notification settings - Fork 146
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
Instance + Broker Pod disappears #372
Comments
I think there are two issues happening here:
|
👍🏻 I have zero rust knowledge, but I think for issue #1 (which I think in large part would stabilize my usage of Akri), the area of the code we'd be looking at is here. So I guess the first thing is we need to be able to check to see what changed. And if its just metadata or status, then we ignore the change, otherwise do what we do today. |
@ammmze, i agree that's where the action needs to happen. I dont think there is a simple way of checking if metadata or status changed, since we aren't storing the previous Configuration. We could add that to the It's not super trivial, but would you be interested in putting in the fix? Trying out rust? |
👍🏻 Yea I am up for giving it a go. Probably won't happen super quick since it'll happen between work and family life stuff, but it seems reasonably simple for a intro to rust 😀 |
For the race condition issue, maybe this doesn't do what I'm thinking (again still very much a noob with rust ... I come from primarily java and javascript), but this seems odd to me that we are cloning the config map when we do the delete. Then the add happens on the original config_map. I'm probably missing something or not fully understanding it, but just figured I'd toss it out there since I'm looking at this thing. |
That is cloning a reference to the shared map which is of type |
First part fix in PR #373 |
I did some investigation into the instances not being recreated, and like you mention in #373, i think this started after |
Awesome! Thank you! Excited to see these changes 🎉 ! |
Describe the bug
I'm having an issue where my Instance and it is associated Broker pod seem to disappear and won't come back until I either restart my agent and udev discovery pods, or it seems I can simply modify my
Configuration
resource which triggers it to come back.Output of
kubectl get pods,akrii,akric -o wide
Kubernetes Version: 1.21.3
Logs
I believe these are applicable logs from the controller. I just had my
conbee-ii
instance disappear on me. The discussion we had in Slack points to potential race conditions happening when theConfiguration
resources are updated. Typically as a result of Flux updating metadata on the resources to manage its state. So we would likely have multiple resources being updated at nearly the same time.The text was updated successfully, but these errors were encountered: