Redis on RPi (arm64) enters in CrashLoopBackOff state
I'm trying to run nextcloud on a raspberry pi cluster (k3s) I've successfully installed nextcloud by removing the bitnami images and using the official ones.
However, Nextcloud's performance was not optimal, so I tried to enable redis:
redis:
enabled: true
image:
registry: docker.io
repository: redis
tag: 6.2.1-buster
usePassword: false
password: "changeme"
global:
storageClass: nfs-client
Unfortunately, no matter what I do, the pods enter in a CrashLoopBackOff state. The only logs I can get are:
[I] ➜ kubectl logs nextcloud-redis-slave-0
/opt/bitnami/scripts/start-scripts/start-slave.sh: line 21: /run.sh: No such file or directory
[I] ➜ kubectl logs nextcloud-redis-master-0
/opt/bitnami/scripts/start-scripts/start-master.sh: line 21: /run.sh: No such file or directory
According to the pod's events, is correctly downloading the image from the official redis repository. Has anyone figured out a solution for this problem?
Hi,
bitnami doesn't support arm64. I had the same problem with postgresql, you just can use the official redis image. Just put the following code into your values.yaml file:
redis:
image:
repository: redis
tag: 6.2.2
This should solve your problem.
@steled If you notice my script, I'm not really using the bitnami image. (Although the file path in logs point to that)
Sorry, for that.
But what looks strange to me is the path in the logs: /opt/bitnami/scripts/start-scripts/start-master.sh
The bitnami folder is a bit strange there I think.
I have to agree on that point.
If you check the yaml of the deployment, the spec looks like this:
spec:
containers:
- command:
- /bin/bash
- -c
- /opt/bitnami/scripts/start-scripts/start-master.sh
So it seems to be somewhat hardcoded to require a bitnami image.
Based on the Bitnami Chart Values You could override the command using the master.command value.
Bitnami now supports arm images, so this should be good to go as of this bitnami/charts#7305. If you're still having trouble, please anonymize your values.yaml of any sensitive data and post it in full here, and we can help take a look.
I will close this issue due to inactivity probably in another month or two.
Closing as stale due to no activity, but as always, feel free to open a new issue if you or others in the community are still having an issue with the latest version of this helm chart. 🙏