Init Container returns with a non-zero causing the helm chart to fail

[jstralko]

What happened:

When installing the csi-blob helm chart on a new 1.27.3 Kubernetes cluster the csi-blob daemonset fail to start. The pods fail to start and immediately go into Init:CrashLoopBackOff state. Digging into the init container the following script returns a non-zero: /blobfuse-proxy/init.sh

Here's the output:

The following packages will be DOWNGRADED:
blobfuse2
0 upgraded, 0 newly installed, 1 downgraded, 0 to remove and 55 not upgraded.
E: Packages were downgraded and -y was used without --allow-downgrades.

What you expected to happen:

I expect the init container to return a zero so the pods can get in Running state.

Environment:
kubectl get po -o yaml | grep mcr | grep blob
image: mcr.microsoft.com/oss/kubernetes-csi/blob-csi:v1.23.2
image: mcr.microsoft.com/oss/kubernetes-csi/blob-csi:v1.23.2
imageID: mcr.microsoft.com/oss/kubernetes-csi/blob-csi@sha256:31854ed041f1d59f8210976fa2b71e39b0628d9d3ff868a46d444d7fcf401fb7

• CSI Driver version: 1.23.2
• Kubernetes version (use kubectl version):
  Client Version: v1.28.3
  Kustomize Version: v5.0.4-0.20230601165947-6ce0bf390ce3
  Server Version: v1.27.3

[jmclong]

I'm sorry that you're experiencing this issue with the csi-blob daemonset.

@jstralko, can you try using the workaround that Andy mentioned here? This should skip installing blobfuse2

#998 (comment)

[jstralko]

Thank you, i will try that now. That is a much better workaround then what i'm doing now. I was patching the initContainer with this:
- command: ["/bin/sh","-c"]
args: ["/blobfuse-proxy/init.sh || true"]

[jstralko]

FYI: to fix my automation (in terraform) i was able to install the helm chart with the following value:
node.blobfuseProxy.installBlobfuse2 to false

[andyzhangx]

if you could update the charts repo, it works now with blobfuse v2.1.2 installed by default