State of ext4 auto_da_alloc-like workarounds in XFS

[Eric Sandeen]

On 12/21/15 12:37 PM, Vallo Kallaste wrote:
> Hi
> 
> The ext4 workaround for wrong-but-common truncate/rename cases can be found at
> https://en.wikipedia.org/wiki/Ext4#Delayed_allocation_and_potential_data_loss
> This is the default "auto_da_alloc" mount option.
> 
> There is no mention of similar workaround in XFS FAQ at
> http://xfs.org/index.php/XFS_FAQ nor
> Wiki article at http://en.wikipedia.org/wiki/XFS
> I suggest adding such a section.
> 
> I have read the thread at
> http://article.gmane.org/gmane.comp.file-systems.xfs.general/53366
> There seems to be definite answer at the top of
> http://article.gmane.org/gmane.comp.file-systems.xfs.general/53376
> 
> But besides that I found
> http://article.gmane.org/gmane.linux.kernel/812588 where it's mentioned that
> /--/
> The bigger problem from my perspective is that XFS is only doing this for the
> truncate case, and (from what I've been told) not for the rename case.
> /--/
> 
> I'd like to know the current state of ext4 auto_da_alloc-like workarounds in XFS,
> particularly for RHEL7. Considering the two cases in
> https://en.wikipedia.org/wiki/Ext4#Delayed_allocation_and_potential_data_loss
> is XFS behaving the same as ext4, both mounted with default options?

The sync-on-close-after-file-got-truncated case has been handled since 2007; see

https://git.kernel.org/cgit/linux/kernel/git/dgc/linux-xfs.git/commit/?id=ba87ea699ebd9dd577bf055ebc4a98200e337542

The sync-after-rename behavior was suggested and rejected for xfs, see

http://marc.info/?t=139845506300002&r=1&w=2

If you'd like to add this information to the XFS wiki, please do so!

Thanks,
-Eric