Open mstingl opened 2 days ago
Hi thank you for the issue 👍
This happens when the job resource is modified from other operators or in some way. That's mostly true if you see the diff hash log/event.
Update Store my-shop setup job in namespace my-shop-namespace. Diff hash
If the job resource is not the same as it was before, we delete the job and recreate it. Because jobs can't be patched and so we delete and recreate it.
The first issue
job not found in container: shopware-setup
happens when the setup job is found but the name of the setup container inside the job is not shopware-setup. You can't modify the name, because it's hard coded in the code so I guess the issue is a timing issue with the api and the operator.
We can fix this in the operator by ignoring fields for the hash generation. So if you can tell me what get's attached after the job is created, I can fix it. Annotations should be fine, but labels for example would trigger a recreate of the job.
Steps to reproduce
Deployment of a Shop through shopware helm chart
What should happened?
The setup job should pass
What actually happened?
The setup job is instantly deleted (and the pod killed) after creation, leaving the operator stuck in the setup stage. Repeating every 10 seconds
Relevant log output
(Namespace, deployment name and URLs are changed)
Your custom resource