Corrupt VG after snapshot backup

M@ario

Member
Dec 23, 2010
34
0
6
Hello!

I have a 5 node proxmox cluster with a shared FC Storage over Device Mapper and a single VG. I only use KVM VMs.
Recently I activated the backup on all nodes witch snapshots. Everythiong went fine for a while but now I had various LVM related errors.
Checksum errors on device-mapper devices and so on.

After a reboot of all nodes I had to fix the VG with vgcfgrestore. Everything seems to be fine again but there are snapshot LVM volumes and LVM volumes with -real and -cow extensions all over the place
brw-rw---- 1 root disk 251, 9 15. Sep 09:43 VM_Data-vm--41872--disk--1-real
brw-rw---- 1 root disk 251, 19 15. Sep 09:43 VM_Data-vzsnap--grzitovirt01e--1-cow

Can I safely delete (lvremove) the vzsnap snapshots? How about the device nodes? I see them on all cluster nodes. Do they get removed automatically?
What about those -real devices? What are they and how can I get rid of them?