Cannot delete this raw disk from Ceph storage pool

[Ahmad Dhamiri]

Greetings all.

This is my first time playing around Proxmox environment. Yesterday, I was trying to redo a Ceph pool storage and try to remove a container from the cluster that is connected to the Ceph pool storage. The container is a success, but then I encounter this problem where I cannot delete this raw disk file. I've tried several methods I can think of, but even in command line I can't seem to find this disk.

Anyone have the solution?

I'm using Proxmox VE 5.4 by the way.

 

[Alwin]

It seems the disk is not on the Ceph pool anymore. Try a 'rbd -p <pool_name> ls' to see if it still exists there.

 

[Ahmad Dhamiri]

Thanks for the response Alwin.

I just issue the command you've mentioned and it shows that the disk is still available.
Is there any command I need to issue to remove the disk?

Refer to the figure below.



Best regards,
Ahmad Dhamiri.

 

[Alwin]

Does the template or the container still exist? If so, then you need to delete them or remove the disk from their config.

 

[Ahmad Dhamiri]

I'm afraid I've deleted the template. Tried to ls several times but to no avail.
Is the location in the figure below to the template is correct?



Regards,
Ahmad Dhamiri

 

[Alwin]

Ok, if you are sure, you don't need it anymore, then run a 'rbd -p <pool> rm base-101-disk-0' to remove it from ceph.

 

[Ahmad Dhamiri]

Hi Alwin,

Sorry for the late reply. Just tried to run the command you suggested, I think it's almost success but then it issues this alert (refer to the image below):


Here are questions popped in my head right now:
1) Is watcher mentioned in the error message means the deleted container?
2) If so, how do I unmap the image from the already deleted watcher?
3) In future, if I want to delete a container, should I remove the image (disk) attached to it first?

Regards,
Ahmad Dhamiri

 

[Alwin]

1) Is watcher mentioned in the error message means the deleted container?

yes.

2) If so, how do I unmap the image from the already deleted watcher?

'rbd unmap /dev/rbd/<image>', but the watcher just means that a connection is holding the image open, not that there is an acting client.

3) In future, if I want to delete a container, should I remove the image (disk) attached to it first?

This is usually done by the removal of the container itself. Whatever happend during the container removal, stopped the deletion of the image.

 

[Ahmad Dhamiri]

1) yes.
2) 'rbd unmap /dev/rbd/<image>', but the watcher just means that a connection is holding the image open, not that there is an acting client.
3) This is usually done by the removal of the container itself. Whatever happend during the container removal, stopped the deletion of the image.

Thank you for the answers Alwin, appreciate the feedback.

I think there is something happened during the deletion of the VM/Container, but I couldn't trace it.
I think I might skip this problem for now since it dosen't affect my playthrough, I hope.