Andy Zhang
Andy Zhang
@ejschoen yes, you need to update the secret referenced by csi.nodeStateSecretRef and then restart pod to make volume remount happen.
the mount process looks good, if you could ssh to the agent node, there is a standalone `blobfuse2` process running for a single mount (see below example), the blobfuse2 would...
then you could run "mount | grep nfs" to find out the nfs mount on the node, and check whether it's read-only on the node directly
hi @rainerleber `/usr` dir is set as readonly by gardener, right? in the CSI driver setting, we don't set the `/usr` dir as readonly during volume mount.
could you set `--node.enableBlobfuseProxy=false` in helm chart install or set `INSTALL_BLOBFUSE_PROXY` env as false in driver daemonset on the node? and what's the possible install path in gardener ?
@rainerleber is `/usr/local/bin` path writable in gardener node?