[PVE-User] Again trouble, but this time with ext4/trim...

[Fabian Grünbichler]

On Mon, Jul 29, 2019 at 03:36:38PM +0200, Marco Gaiarin wrote:
> 
> > In that servers i've also some other FS, but the ext4 ones are low
> > varying, mounted RO or noatime, but also some FS in XFS fs, that seems
> > does not suffer.
> 
> I've disabled 'discard' for /dev/sda in both server, and keeped it for
> /dev/sdb (squid spool, ext4) and /dev/sdc (XFS filesystem).
> 
> I've also added a 'randomization' for the fstrim task, but still:
> 
>  Jul 29 01:14:10 vdmsv1 systemd[1]: Starting Discard unused blocks...
>  Jul 29 01:14:14 vdmsv1 kernel: [455671.833357] EXT4-fs error (device sdb1): ext4_validate_block_bitmap:386: comm fstrim: bg 97: bad block bitmap checksum
>  Jul 29 01:14:14 vdmsv1 kernel: [455671.859823] EXT4-fs warning (device sdb1): ext4_trim_all_free:5130: Error -74 loading buddy information for 97
>  Jul 29 01:14:14 vdmsv1 fstrim[23986]: fstrim: /var/spool/squid: FITRIM ioctl failed: Messaggio errato
>  Jul 29 01:15:02 vdmsv1 CRON[24072]: (root) CMD (command -v debian-sa1 > /dev/null && debian-sa1 1 1)
>  Jul 29 01:15:29 vdmsv1 systemd[1]: fstrim.service: Main process exited, code=exited, status=64/n/a
>  Jul 29 01:15:29 vdmsv1 systemd[1]: Failed to start Discard unused blocks.
>  Jul 29 01:15:29 vdmsv1 systemd[1]: fstrim.timer: Adding 2h 8min 41.708255s random time.
>  Jul 29 01:15:29 vdmsv1 systemd[1]: fstrim.service: Unit entered failed state.
>  Jul 29 01:15:29 vdmsv1 systemd[1]: fstrim.timer: Adding 3h 56min 32.453953s random time.
>  Jul 29 01:15:29 vdmsv1 systemd[1]: fstrim.service: Failed with result 'exit-code'.
> 
> So seems that i've some trouble with 'discard', at least with ext4
> FS...

are you sure this storage box supports fstrim at all? thin provisioning
and trim/discard often go together, but they don't have to..

quick googling only showed some vmware specific integration to 'reclaim
no longer used space of thin provisioned volumes', but no explicit
mention of discard support.