[dm-devel] trouble with generic/081

[Eric Sandeen]

On 1/5/17 3:12 PM, Zdenek Kabelac wrote:
> Dne 5.1.2017 v 20:29 Eric Sandeen napsal(a):
> 
> So  if I hear all voice correctly now -
> 
> 
> we now want to let user continue to use such systems and let them
> figure out themself something is wrong when they get occasional write
> failure and XFS now avoid destruction by shutting down on any journal
> failure.

No - I'd like to know, specifically, what scenario leads to filesystem
corruption when we run out of space in the thin-pool, because I do
not expect that result.

Anecdotes about "destruction" don't really help.

>> (A journal may not be replayable on mount if it needs to allocate more
>> thin blocks on replay and is unable to do so, but hat should just fail
>> gracefully)
> 
> 
> I don't have a test sample myself - just some guides how to get to it.
> 
> Use  LV and make some thin snapshots.
> 
> Then change various parts of origin - at various moment before pool is out-of-space
> 
> So you will get lots of different scenarios of missing data.

(missing /data/ is not "filesystem destruction" - if we're just talking
about buffered IO which was never fsynced and does not persist on
disk, that is expected with or without thinp on any sort of storage
trouble.)
 
> You will mostly not get into those mentioned trouble if you
> have just single thinLV and you exhaust thin-pool while using it.
> 
> Games with snapshot are needed.

Please formalize that, and I'll be happy to take a look.

-Eric