On Tue, Apr 21, 2015 at 3:43 PM, Ilya Dryomov <idryomov@xxxxxxxxx> wrote: > On Tue, Apr 21, 2015 at 11:49 AM, Stefan Priebe - Profihost AG > <s.priebe@xxxxxxxxxxxx> wrote: >> Hi, >> >> while running firefly i've seen that each osd log prints: >> 2015-04-21 10:24:49.498048 7fa9e925d780 0 >> xfsfilestorebackend(/ceph/osd.49/) detect_feature: extsize is disabled >> by conf >> >> The firefly release notes V0.80.4: >> osd: disable XFS extsize hint by default (#8830, Samuel Just) >> >> But hammer also mentions: >> osd: verify kernel is new enough before using XFS extsize ioctl, enable >> by default (#9956 Sage Weil) >> >> I'm now confused. >> >> Should i enable this one? Or keep it disabled? What's the advantage / >> disadvantage? > > As far as I know the story there is as follows. > > extsize was added to try to do a little better on the fragmentation > front, especially in rbd workloads. Shortly after it was added and > enabled by default we hit a bug in xfs and disabled the extsize hint. > After some back and forth with xfs maintainer and a lot of testing by > Sam we found out that the bug only affects < 3.5 kernels. Following that > a safeguard was added to XfsFileStoreBackend so that you can enable > extsize hinting only if you are running a >= 3.5 kernel. > > It is now enabled by default so I'd say you should enable it. Here is the link to that bug: http://oss.sgi.com/bugzilla/show_bug.cgi?id=874 Thanks, Ilya _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |