mirror of
https://github.com/pocket-id/pocket-id.git
synced 2025-12-09 17:23:22 +03:00
🐛 Bug Report: Docker image for version v1.2.0 fails to start #178
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @beauwest on GitHub.
Reproduction steps
When I updated the tag on my image, I get an error. Possibly the same issue as #208 .
Expected behavior
It starts
Actual Behavior
Version and Environment
ghcr.io/pocket-id/pocket-id:v1.2.0
Log Output
No response
@beauwest commented on GitHub:
Previous version was v1.1.0.
OS is Talos Linux so running the container in a pod on Kubernetes.
CPU architecture is x86-64.
@beauwest commented on GitHub:
Since it's in a Kubernetes environment, there isn't a docker CLI. Here's the output of
kubectl describe pod …And only output in the pod log
@stonith404 commented on GitHub:
Thanks but it seems like this doesn't output the ARCH of the image. We are not officially supporting the Kubernetes installation so I can't really help you here if the issue is related to Kubernetes.
Are you able to run the same image on the same machine without Kubernetes?
@stonith404 commented on GitHub:
Thanks, can you share the output of
docker image inspect ghcr.io/pocket-id/pocket-id:v1.2.0. You explicitly use the image tagv1.2.0, right?@stonith404 commented on GitHub:
Which was your previous version and which OS and CPU architecture are you using?
@beauwest commented on GitHub:
While I can't rule it out, my gut feeling would be that it's not an issue specifically related to Kubernetes, especially since the v1.1.0 works fine. Unless something significant changed in the entrypoint.sh file?
@beauwest commented on GitHub:
For future travelers. This ended up being some kind of issue with a particular node in the Kubernetes cluster. Other nodes could run the 1.2.0 image correctly, and deleting and recreating the node itself solved the issue. The clue was when other containers started failing as well.
Not a pocket-id issue!
@beauwest commented on GitHub:
Unfortunately that's not testable. Talos is an immutable OS specifically for Kubernetes. I did test a simple docker configuration on a different machine, with an arm64 architecture, and arm64 v1.2.0 does work at least. I'm unable to test amd64.
@stonith404 commented on GitHub:
I also thought that's probably not related to Kubernetes but I was able to run
v1.2.0on a amd64 machine with Docker. Strange is also that we didn't change anything related to Docker in the latest release, you can see all changed files here.I'm closing this issue as this only seems to occur with Kubernetes but feel free to reopen the issue if you can reproduce it with an official installation method.