Anyone here using a custom flexvol driver like this:
We have pulled all out hair out trying to get this to work on any version of 1.10.x or 1.11.2.
We’ve gone as far as rewritting this in python and added heavy error checking and debug code. We’ve added support for master initiated attach/detach = false.
We can’t get either mode to work.
Supposedly when master attach/detach is set to false, the node does the operations and the driver should only support “init”, “mount” and “unmount”.
Yet it still tries to call getvolumename(). If we add a stub function to return a hard-coded volume name it get’s passed this error and we reach the current roadblock we can’t get around:
2018-09-06T07:06:28.570528-07:00 st11p01if-ztds07071001 kubelet[19142]: W0906 07:06:28.570254 19142 driver-call.go:144] FlexVolume: driver call failed: executable: /usr/libexec/kubernetes/kubelet-plugins/volume/exec/bithub~lvm/lvm, args: [attach {“kubernetes.io/fsType":“ext4”,“kubernetes.io/pvOrVolumeName”:“test”,“kubernetes.io/readwrite”:“rw”,“ls_fsopts”:"",“lv_fstype”:“ext4”,“lv_name”:“testvol”,“lv_size”:“10G”,“mountopts”:“rw,noatime”,“mountpoint”:"/data",“vg_name”:"vg1”} st11p01if-ztds07071001.example.com], error: exit status 1, output: “{“status”: “Not Supported”}\n”
2018-09-06T07:06:28.571636-07:00 st11p01if-ztds07071001 kubelet[19142]: E0906 07:06:28.570317 19142 nestedpendingoperations.go:267] Operation for ““flexvolume-bithub/lvm/test”” failed. No retries permitted until 2018-09-06 07:08:30.570285576 -0700 PDT m=+513910.739491088 (durationBeforeRetry 2m2s). Error: "AttachVolume.Attach failed for volume “test” (UniqueName: “flexvolume-bithub/lvm/test”) from node “st11p01if-ztds07071001.example.com” : attach command failed, status: Not Supported, reason: "
We’ve submitted bugs against the 1.10.2 and 1.10.6 versions and they’ve received no attention what-so-ever.