Cannot start VM - /dev/rbd/rbd missing, but /dev/rbd0 present?

[jw6677]

/dev/rbd1
kvm: -drive file=/dev/rbd/rbd/vm-150-disk-0,if=none,id=drive-scsi0,discard=on,format=raw,cache=none,aio=io_uring,detect-zeroes=unmap: Could not open '/dev/rbd/rbd/vm-150-disk-0': No such file or directory
TASK ERROR: start failed: QEMU exited with code 1

I suspect this issue has resulted from a recent accidental upgrade to ceph, outside of the pve repo.

It seems I can `rbd map` drives to /dev/rbdX properly, but when attemping to start a VM, I receive the above error.

Upon inspection the `/dev/rbd` directory does not exist.

Can anyone offer guidance as to what I should read/learn about, in order to solve this issue?


Edit: In related news, Manually creating via `mkdir /dev/rbd/rbd && ln -s /dev/rbd/rbd/vm-150-disk-0 /dev/rbd1` appears to allow the machine to start, so I have my workaround and am closer to the solution.
It seems to me there is a system for controlling the creating of links (etc) when devices are added to a machine, but I am a little lost...

 

[jw6677]

Well that was super tedious to figure out.

For anyone who comes in future, my issue was with `/usr/bin/ceph-rbdnamer` which was missing for some stupid (User error) reason.

udev is the system I was trying to identify.

Other possible causes for this issue would be:
Issues with user udev rules at: `/usr/lib/udev/rules.d/`
Issues with this file in particular: `/usr/lib/udev/rules.d/50-rbd.rules`