On Monday, June 17, 2019 at 11:32:50 AM UTC-4, Chris Laprise wrote:
> FWIW, if there were any issues with data not being discarded, it would 
> be with the (size) mismatch between what ext4 considers a discarded 
> block and what the thin + lvm layers consider a discardable block or chunk.

Yes, this is why turning on discards in fstab inside VM templates really only 
impacts large file deletes inside the VMs, as the pool chunk sizes is >= 64KB 
while the VM block size inside the VMs is much smaller. Ensuring that one 
deletes all of the directories/files (big and small) *then* performing fstrim 
before shutdown can partially address this.

Also having Qubes tooling explicitly issue a blkdiscard against devices 
*before* performing lvremove partially addresses this as well (for deleted or 
disposable VMs), for the persistent LVs.

Partial solutions aren't great for this topic, though.

> If I wanted to solve this issue relatively quickly, I'd first consider 
> moving to btrfs. Its unified approach is more likely to process discards 
> completely.

I keep getting mixed messages on the stability of btrfs. :)

Brendan

-- 
You received this message because you are subscribed to the Google Groups 
"qubes-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/qubes-users/3fdd7eee-f2d1-409c-8aa7-1d1379418649%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to