[Gluster-users] Perfomance issue on a 90+% full file system

[Dan Mons]

On 8 October 2014 09:40, Franco Broi <franco.broi at iongeo.com> wrote:
> ZOL isn't the fastest filesystem out there, so maybe coming off a lower
> base level hides the impact of fragmentation somewhat.

That was my first thought.  Although it might just be a ZFS feature
too, given it's clever/modern design.

> What are you using for RAID? Judging from the video on your website, we
> are using similar hardware.

LSI 9260-8i cards in all of our nodes currently.  16x 3TB Hitachi SAS
drives in RAID6+1S.

> Not sure where the 90% figure comes from, surely it depends on the disk
> capacity???

90% is anecdotal based on history (the last ~10 years or so of my
sysadminning).  I originally thought it would be down to a hard value
free, but the percentage appears to scale with storage (at least, so
far it does - there may well be an upper limit, although by then I
hope I'm free of legacy file systems like XFS).

I'm also interested to hear of anyone using BtrFS, and if these
problems exist there too (either in standard non-GlusterFS NAS units,
or GlusterFS setups).

-Dan