[linux-lvm] on discards

[Xen]

On discards,

I have a thin pool that filled up when there was enough space, because
the filesystem hadn't issued discards.

I have ext4 mounted with the discard option, I hope, because it is in
the list of default mount options of tune2fs:

Default mount options: user_xattr acl discard

What can I do to ensure that it won't happen again?

Is the filesystem really mounted with discard?

It doesn't say so anywhere.

Regards.

[Zdenek Kabelac]

Post by Xen
On discards,
I have a thin pool that filled up when there was enough space, because the
filesystem hadn't issued discards.
I have ext4 mounted with the discard option, I hope, because it is in the list
Default mount options: user_xattr acl discard
What can I do to ensure that it won't happen again?
Is the filesystem really mounted with discard?
It doesn't say so anywhere.

Hi


I'd not suggest to mount fs with discard - but rather run 'fstrim' in your
cron once a while.

Standalone 'small trim' on deleted file is quite inefficient and will not
resolve aggregation of continuous blocks for proper trimming.

For thin-pool - min size of trim is i.e. whole 64K chunk, - when it sits on
SSD you need to even trim much multiple pool chunks to free something on your SSD.

When you run 'fstrim' whole region of unused blocks is trimmed - to it's
always biggest possible chunk.

Regards

Zdenek