Description
/kind bug
Hello everyone
What happened:
I'm evaluating openstack-cinder-csi as a CSI provider for my kubernetes cluster and I couldn't resize my PVC.
When looking at the openstack-cinder-csi-nodeplugin pod logs, I found the following error message :
I0404 09:45:04.087521 12 mount_linux.go:542] Checking for issues with fsck on disk: /dev/disk/by-id/scsi-0QEMU_QEMU_HARDDISK_eb7b56ee-343e-40a9-9aa6-313f418d324c
I0404 09:45:04.108998 12 mount_linux.go:555] `fsck` error fsck from util-linux 2.37.2
fsck: /sbin/fsck.xfs: execute failed: No such file or directory
The binary /sbin/fsck.xfs is present on the docker image (see tools/csi-deps-check.sh) but /sbin/fsck.xfs is a shell script asking for /bin/sh :
~ file /sbin/fsck.xfs
/sbin/fsck.xfs: POSIX shell script, ASCII text executable
~ head /sbin/fsck.xfs
#!/bin/sh -f
#
# Copyright (c) 2006 Silicon Graphics, Inc. All Rights Reserved.
#
NAME=$0
# get the right return code for fsck
repair2fsck_code() {
case $1 in
I created a custom debian based docker image (no distroless, with all the bash/sh utilities) and put the cinder-csi-plugin binary on it and didn't have encounter the problem after that.
What you expected to happen:
My PVC to be resized
How to reproduce it:
Create a XFS storage class, a PVC and a deployment using the PVC :
apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
name: csi-cinder-sc-delete-netapp-xfs
allowVolumeExpansion: true
allowedTopologies:
- matchLabelExpressions:
- key: topology.cinder.csi.openstack.org/zone
values:
- DC1
- DC2
parameters:
csi.storage.k8s.io/fstype: xfs
availability: nova
type: XXXX
provisioner: cinder.csi.openstack.org
reclaimPolicy: Delete
volumeBindingMode: Immediate
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: httpbin
namespace: bench-csi
spec:
accessModes:
- ReadWriteOnce
resources:
requests:
storage: 1Gi
storageClassName: csi-cinder-sc-delete-netapp-xfs
---
apiVersion: apps/v1
kind: StatefulSet
metadata:
name: httpbin
namespace: bench-csi
spec:
replicas: 1
selector:
matchLabels:
app: httpbin
template:
metadata:
labels:
app: httpbin
spec:
containers:
- name: httpbin
image: kennethreitz/httpbin
volumeMounts:
- mountPath: /data
name: data-volume
ports:
- containerPort: 80
resources:
requests:
memory: "1Gi"
cpu: "1000m"
limits:
memory: "2Gi"
cpu: "2000m"
livenessProbe:
httpGet:
path: /status/200
port: 80
initialDelaySeconds: 3
periodSeconds: 3
volumes:
- name: data-volume
persistentVolumeClaim:
claimName: httpbinRun the df -h command into the pod :
kubectl exec -it $(kgp -l app=httpbin --no-headers |awk '{print $1}') -- bash -c "df -h |grep data"
/dev/sdb 960M 39M 922M 5% /data
Patch the PVC size and re run the df -h command :
kubectl patch pvc httpbin -p '{"spec":{"resources":{"requests":{"storage":"5Gi"}}}}'
kubectl exec -it $(kgp -l app=httpbin --no-headers |awk '{print $1}') -- bash -c "df -h |grep data"
/dev/sdb 960M 39M 922M 5% /data
Anything else we need to know?:
Environment:
- openstack-cloud-controller-manager(or other related binary) version: 2.32.0
- OpenStack version: Yoga
- Others: