pod "ceph-osd-dev-" blocked after the k8s recovered from a crash

ceph-helm/ceph/ceph/templates/bin/_osd_disk_prepare.sh.tpl

Line 30 in 743a744

mount ${OSD_DEVICE}1 ${tmp_osd_mount}

I installed a ceph cluster using the ceph-helm, but my kubernetes cluster crashed for some reason. After I restored the k8s cluster, all the pods recovered excluding the ceph-osd-dev- pods.

The initContainer osd-prepare-pod logs:

2018-08-25 04:37:41  /start_osd.sh: Checking if it belongs to this cluster
++ echo 2712
+ tmp_osd_mount=/var/lib/ceph/tmp/2712/
+ mkdir -p /var/lib/ceph/tmp/2712/
+ mount /dev/loop01 /var/lib/ceph/tmp/2712/
mount: special device /dev/loop01 does not exist

I checked the devices in /dev/ ,

$ ls /dev/loop*
/dev/loop0  /dev/loop0p1  /dev/loop0p2  /dev/loop-control

I suspect the /dev/loop0p1 is the device wanted, so I make a symbol link , and IT WORKS!

I think there is something wrong with it. The ceph deamon image is

registry.docker-cn.com/ceph/daemon                                         tag-build-master-luminous-ubuntu-16.04   c48fa6936ae5        6 months ago        445 MB