we have a proxmox 8.0.3 running on ceph 17.2.6 . This clsuer is older and updated multiple times.
Today we run the command:
There we find out that in the past there must have been a problem with deleting TPM snapshot disks.
We have a lot of VMs that look slike this. In proxmox there are 10 snapshots in the UI.
Correct disk 0 :
Correct disk 1 :
Incorrcect TPM disk 2
New snapshots are deleted correctly. So it must have been fixed.
Is there a way to delete all the old unused disk snapshots, where there is no corresponding proxmox snapshot anymore ?
My workaround would be, to delete all snapshots from 2022 and next year i could delete all left over disk snapshots from 2023
something ilke
27.10.2023 : The rbd command does not allow wildcards in there snapshot remove command. So we ended up using this pipe
Yes, this is very specific for our problem. But i thought i leave it here, in case someone needs it as inspiration.
Today we run the command:
rbd du --pool <poolname> There we find out that in the past there must have been a problem with deleting TPM snapshot disks.
We have a lot of VMs that look slike this. In proxmox there are 10 snapshots in the UI.
Correct disk 0 :
Code:
vm-100-disk-0@autoweekly231001220413 1 MiB 1 MiB
vm-100-disk-0@autoweekly231008220352 1 MiB 0 B
vm-100-disk-0@autoweekly231015220350 1 MiB 0 B
vm-100-disk-0@autodaily231020220806 1 MiB 1 MiB
vm-100-disk-0@autodaily231021220640 1 MiB 0 B
vm-100-disk-0@autodaily231022220508 1 MiB 0 B
vm-100-disk-0@autoweekly231022220512 1 MiB 0 B
vm-100-disk-0@autodaily231023220001 1 MiB 1 MiB
vm-100-disk-0@autodaily231024220001 1 MiB 0 B
vm-100-disk-0@autodaily231025221734 1 MiB 1 MiB
vm-100-disk-0 1 MiB 0 BCorrect disk 1 :
Code:
vm-100-disk-1@autoweekly231001220413 400 GiB 42 GiB
vm-100-disk-1@autoweekly231008220352 400 GiB 26 GiB
vm-100-disk-1@autoweekly231015220350 400 GiB 31 GiB
vm-100-disk-1@autodaily231020220806 400 GiB 27 GiB
vm-100-disk-1@autodaily231021220640 400 GiB 10 GiB
vm-100-disk-1@autodaily231022220508 400 GiB 9.2 GiB
vm-100-disk-1@autoweekly231022220512 400 GiB 204 MiB
vm-100-disk-1@autodaily231023220001 400 GiB 7.8 GiB
vm-100-disk-1@autodaily231024220001 400 GiB 19 GiB
vm-100-disk-1@autodaily231025221734 400 GiB 21 GiB
vm-100-disk-1 400 GiB 14 GiBIncorrcect TPM disk 2
Code:
vm-100-disk-2@autodaily220510010001 4 MiB 4 MiB
vm-100-disk-2@autodaily220511010448 4 MiB 4 MiB
vm-100-disk-2@autodaily220512010512 4 MiB 0 B
vm-100-disk-2@autodaily220513010418 4 MiB 0 B
vm-100-disk-2@autodaily220514010443 4 MiB 0 B
vm-100-disk-2@autodaily220515010251 4 MiB 0 B
vm-100-disk-2@autodaily220516010518 4 MiB 0 B
vm-100-disk-2@autodaily220517010700 4 MiB 4 MiB
vm-100-disk-2@autodaily220518010641 4 MiB 0 B
vm-100-disk-2@autodaily220519010620 4 MiB 0 B
vm-100-disk-2@autodaily220520010631 4 MiB 0 B
vm-100-disk-2@autodaily220521010616 4 MiB 0 B
vm-100-disk-2@autoweekly220522010350 4 MiB 0 B
vm-100-disk-2@autodaily220522010529 4 MiB 0 B
vm-100-disk-2@autodaily220523010632 4 MiB 0 B
vm-100-disk-2@autodaily220524010622 4 MiB 0 B
vm-100-disk-2@autodaily220525010524 4 MiB 4 MiB
vm-100-disk-2@autodaily220526010217 4 MiB 4 MiB
vm-100-disk-2@autodaily220527010206 4 MiB 0 B
vm-100-disk-2@autodaily220528010207 4 MiB 0 B
vm-100-disk-2@autoweekly220529010125 4 MiB 0 B
vm-100-disk-2@autodaily220529010143 4 MiB 0 B
vm-100-disk-2@autodaily220530010215 4 MiB 0 B
[..manny more entries]New snapshots are deleted correctly. So it must have been fixed.
Is there a way to delete all the old unused disk snapshots, where there is no corresponding proxmox snapshot anymore ?
My workaround would be, to delete all snapshots from 2022 and next year i could delete all left over disk snapshots from 2023
something ilke
rbd snap rm {pool-name}/vm-*@autodaily22*27.10.2023 : The rbd command does not allow wildcards in there snapshot remove command. So we ended up using this pipe
rbd du --pool <poolname> | grep ly22 | awk '{print "<poolname>/"$1}' | xargs -I {} rbd snap rm {}Yes, this is very specific for our problem. But i thought i leave it here, in case someone needs it as inspiration.
Last edited:
