rbd error: rbd: listing images failed: (2) No such file or directory (500)

[Kaboom]

Hi All,

I got this error when I want to see Storage 'ceph_ssd' (Ceph pool name) on node 'node002' (on all nodes).
rbd error: rbd: listing images failed: (2) No such file or directory (500)

When I run rbd ls -l ceph_ssd, I get this at the end:
rbd: listing images failed: (2) No such file or directory

Anyone familiar with this error? Seen some other topics about this, but they all came with a 'vm disk name', but not in my situation.

Thanks!

 

[Alwin]

Might be similar to the below thread, even tough it has an image name in it.
https://forum.proxmox.com/threads/r...failed-2-no-such-file-or-directory-500.56577/

Is there some more information in the logs?

 

[Kaboom]

I don't get any errors about a VM:

....... vm-223-disk-0 50 GiB 2 excl vm-246-disk-0 50 GiB 2 excl vm-247-disk-0 50 GiB 2 excl vm-248-disk-0 50 GiB 2 excl rbd: listing images failed: (2) No such file or directory

 

[Alwin]

Possibly there is a defekt image in that pool. :/

Do you see any entries in the ceph logs? Somewhere it should say if a object is defekt.

 

[Kaboom]

The Ceph logs are looking healthy and no errors or defects. I can see 'summary' but can't see 'content'.

 

[Alwin]

Take the image list, sort it and run through them. It should hang at the image that might have an issue.

 

[Kaboom]

What do you mean with sort it? Stop the VM and run list again?

 

[Kaboom]

I also notice just now when I want to migrate a VM to another node I get this error:

task started by HA resource agent
2020-03-25 21:59:43 starting migration of VM 108 to node 'node004' (10.0.0.1)
2020-03-25 21:59:43 ERROR: Failed to sync data - rbd error: rbd: listing images failed: (2) No such file or directory
2020-03-25 21:59:43 aborting phase 1 - cleanup resources
2020-03-25 21:59:43 ERROR: migration aborted (duration 00:00:00): Failed to sync data - rbd error: rbd: listing images failed: (2) No such file or directory
TASK ERROR: migration aborted

When I want to migrate a container to another node I have no (!) problems.

 

[Alwin]

When I want to migrate a container to another node I have no (!) problems.

Is the container storage a different pool?

What do you mean with sort it? Stop the VM and run list again?

No. Its one liner. Since rbd ls works, you get its output. Sort it and run the rbd info <image> in a loop through the sorted list. The command will hang when it encounters the faulty image.

 

[Kaboom]

Thanks, that did the trick! I found the image and deleted it.

rbd rm vm-xxx-disk-1 -p cephpool

 

[Kaboom]

And the migration also works again! GREAT

 

[Lephisto]

Out of curiosity: how can an rbd image become "faulty" if there was no power or hardware related crash?

 

[Alwin]

Out of curiosity: how can an rbd image become "faulty" if there was no power or hardware related crash?

Hard to say afterwards. But as it's a network distributed storage, a network fault may do that. Or the VM was killed during a process that wrote the header of the RBD image.