There is a new restriction that pool names can't start with a number that I somehow missed in the release notes (although it is there after checking)... We have several pools such as 2hour (for vms to be backed up every 2 hours), 1d6pm and 1d6am for daily backups at obvious times, etc... What's going to happen when these clusters are upgraded? Will the be grandfathered in, or dropped, or...?
Pool names restrictions different in 9
- Thread starter jlauro
- Start date
The easiest thing to do is to install Virtual PVE and test out your exact scenario : digit-named-pool with VM which has disks on it. Start with 8, upgrade to 9.
A quick test on PVE9 shows that the storage subsystem will skip newly incompatible pools, and likely your VMs that have disks on them will fail to start. I have not tested what happens to the VMs that happen to be running at the time of the upgrade, but I suspect they will continue to run until cold start or migration attempt.
Cheers
Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox
A quick test on PVE9 shows that the storage subsystem will skip newly incompatible pools, and likely your VMs that have disks on them will fail to start. I have not tested what happens to the VMs that happen to be running at the time of the upgrade, but I suspect they will continue to run until cold start or migration attempt.
Code:
root@pve-1:/etc/pve# pvesm status
file /etc/pve/storage.cfg line 49 (skip section '2bb-nvme'): storage ID '2bb-nvme' contains illegal characters
Name Type Status Total (KiB) Used (KiB) Available (KiB) %
bb-iscsi blockbridge active 64424509440 4526905536 59897603904 7.03%
bb-nvme blockbridge active 64424509440 4526905536 59897603904 7.03%
root@pve-1:/etc/pve# pvesm alloc 2bb-nvme 100 vm-100-disk-0 1
400 Parameter verification failed.
storage: invalid format - storage ID '2bb-nvme' contains illegal characters
pvesm alloc <storage> <vmid> <filename> <size> [OPTIONS]
root@pve-1:/etc/pve#Cheers
Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox
The place I am using them is resource pools under permissions so they can be used as the filter for backup jobs. They don't show under pvesm status from what I can tell.
Makes sense, you did not specify the type of pool (resource vs storage vs) and I made an assumption of storage as they have similar restriction. My apologies. I don't have sufficient experience with Resource pools. Generally, until a PVE team member chimes in, your best bet is to try this in the lab, or review the related code changes.
Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox
Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox
Last edited:
You can rename the zpool on the fly at import, but then will probably have to update the Storage definitions in Datacenter
zpool import -d /dev/disk/by-id badpoolname newpoolname
zpool import -d /dev/disk/by-id badpoolname newpoolname
If we go back to Storage Pools, which is not Ops question as we know now, one would have to edit/modify every VM configuration file to point to appropriate Pool/Volume.
I think the Resource Pool rename is simpler, one would only need to edit user.cfg.
Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox