Re: [PATCH manager/storage v6 0/2] fix #7339: lvmthick: add option to free storage for deleted VMs

["Michael Köppl" <m.koeppl@proxmox.com>]

Gave this a spin. Used a LUN provided by a SAN via ISCSI with the
underlying storage being thin-provisioned. Then created an LVM-thick
storage with discard enabled, created a new VM with a disk on said
storage and then checked that when removing the VM, the disk space was
reclaimed. Used 
`watch -n 0.5 'lvs -a -o lv_name,data_percent,metadata_percent vgsan'`
to see if the storage was reclaimed as expected.

I also tested what happens if discard is activated and the underlying
storage does not support it. It's obvious from the code that a warning
is printed and then the function carries on, but I think the warning
message could make it more clear what the consequences of the failed
`blkdiscard` command are because right now the warning that is printed
is this:
    command '/sbin/blkdiscard /dev/<vg>/<vol>' failed: exit code 2
I noted on the pve-storage patch that IMO this would benefit from a more
explicit "Disk space could not be reclaimed with discard" or something
similar.

Also had a closer look at the implementation, left some comments on the
individual patches. I also noticed that this lacks a pve-docs patch for
the new discard option.

Consider this:
Tested-by: Michael Köppl <m.koeppl@proxmox.com>

On Fri May 22, 2026 at 3:04 PM CEST, Lukas Sichert wrote:
> Logical volumes (LV) in an LVM (thick) volume group (VG) are
> thick-provisioned, but the underlying backing storage can be

[snip]