Re: ENOSPC design issues

[Josef Bacik <jbacik@fusionio.com>]

On Wed, Sep 26, 2012 at 01:55:47AM -0600, Ahmet Inan wrote:
> when testing, please also do something like this:
> 
> # create big squashfs image somewhere:
> # mksquashfs / /big.img -noappend -no-sparse -e big.img
> 
> # then unpack into fresh filesystem with (and no) compression:
> # unsquashfs -f -d /subvol /big.img
> 
> this is how i was always able to trigger ENOSPC while trying to
> make a full system installation from squashfs image.
> 
> you should also try different compression algos (i only use lzo)
> 
> btw: i was able to trigger ENOSPC with for-linus on 3.5.4 on a
> i686 Pentium M Notebook with only 1GB of Memory and
> fresh FSthis way, otherwise havent seen ENOSPC for long time.
>

Have you actually seen this problem on 3.5 with no compression?  I fixed a
problem in the 3.5 timeframe that should have fixed this for no-compression, and
then I've since fixed the compression part of it in btrfs-next.  Can you retest
with btrfs-next and see if you still see the problem?  In the meantime I'm
running mksquashfs, but it's going to take a few years to complete ;).  Thanks,

Josef