rbd: listing images failed: (2) No such file or directory

[szucs10]

Hi,

Older cluster installation with ceph. At that time, the ceph pool was not created, so the disks were labeled .mgr, but its name appears as a separate storage.
The following error was visible on ceph:

Module 'devicehealth' has failed: disk I/O error

I solved this by regenerating the manager. However, before generating again, one of the 3 managers did not have the usual name (mgr.prod1), but had localhost in its name.

Since the regeneration, I have not been able to access the ceph storage, the VMs are running and there is an I/O operation, but not, for example, I get an error for the following command:

root@prod1:~# rbd -p 'cephname' du
rbd: error opening pool 'cephname': (2) No such file or directory

Does anyone have any idea how I can access the images?

Info for the cluster in the attachment.

Thanks for the help!

b

 

[VictorSTS]

This usually happens when there's a half removed image in a Ceph storage: you've cancelled a clone or a disk move to the Ceph storage. PVE does not fully remove the half copied disk. Dunno if it still happens on more recent versions of PVE/Ceph, but it definitelly happens with PVE 7.2.

List the erroneous images with rbd ls -l <CEPH_POOL_NAME> > /dev/null.
You will see something like:

rbd: error opening vm-506-disk-0: (2) No such file or directory
rbd: listing images failed: (2) No such file or directory

Make sure that the image isn't in use in any VM (grep -R "vm-506-disk-0" /etc/pve/nodes/*/qemu-server/*).
You can remove the RBD image with rbd -p <CEPH_POOL_NAME> rm vm-506-disk-0.