2.6.37: mount with quota eats all available memory

[Arkadiusz Miskiewicz]

On Friday 11 of March 2011, Arkadiusz Miskiewicz wrote:
> On Friday 11 of March 2011, Arkadiusz Miskiewicz wrote:
> > I have a machine with 4GB, running 64bit 2.6.37, xfs on top of soft
> > raid5.
> > 
> > Recently after using xfs_fsr (and getting a oops) had to do xfs_repair
> > and now I'm no longer able to mount this filesystem with usrquota and
> > grpquota.
> 
> Some data about fs:
> # LC_ALL=C xfs_info /home
> meta-data=/dev/md3               isize=256    agcount=32, agsize=6564128
> blks =                       sectsz=4096  attr=0
> data     =                       bsize=4096   blocks=210051776, imaxpct=25
>          =                       sunit=16     swidth=48 blks
> naming   =version 2              bsize=4096   ascii-ci=0
> log      =internal               bsize=4096   blocks=32768, version=2
>          =                       sectsz=4096  sunit=1 blks, lazy-count=0
> realtime =none                   extsz=196608 blocks=0, rtextents=0
> # cat /proc/mdstat
> Personalities : [raid10] [raid1] [raid6] [raid5] [raid4]
> md3 : active raid5 sda4[0] sdd4[3] sdc4[2] sdb4[1]
>       840207168 blocks level 5, 64k chunk, algorithm 2 [4/4] [UUUU]
>       bitmap: 2/3 pages [8KB], 65536KB chunk

More info, tested current linus 2.6.38git including "xfs: Reduce OOM kill 
problems under heavy load V2" patchset but the result wasn't much better:

3549921 3549921 100%    1.06K 132791       30   4249312K xfs_inode
(from slabtop), machine became unresponsible and rebooted.

I wonder if 8GB of ram would be enough for this 800GB fs to mount.
-- 
Arkadiusz Miśkiewicz        PLD/Linux Team
arekm / maven.pl            http://ftp.pld-linux.org/