On 10/26/12 12:15 PM, Milan Holzäpfel wrote:
> Hello all,
>
> I have an XFS filesystem of size 1.2 TiB with 101 GiB free space and 14
> million inodes in use. Quotas are disabled on the filesystem. I try to
> enable user and group quotas by mounting with
>
> mount /dev/xfs_device /mnt -o uquota,gquota
>
> I first tried this with the 3.2.0 Kernel from Ubuntu 12.04 and hit an
> OOM [1, 2]. In [1], it is said that the OOM is fixed in 3.5, so I tried
> 3.5.7 and 3.6.3 from http://kernel.ubuntu.com/~kernel-ppa/mainline/ (I
> believe these are vanilla kernel.org kernels compiled with Ubuntu's
> configuration).
>
> With 3.5.7 and 3.6.3, the OOM does not occur. For some time after
> issuing the mount command, there is much I/O activity with xfs_inode
> in /proc/slabinfo growing up to ~ 500k entries, which is just fine. At
> some point, the I/O activity stops, but the mount does not terminate.
> The system is completely idle. (The mount was the only activity on the
> system.)
>
> In dmesg, I find
>
> INFO: task mount:8806 blocked for more than 120 seconds.
And then what? Probably a backtrace, right?
sysrq-w to get hung tasks or sysrq-t to get all task traces might
help.
The sysrqs are one of the things suggested in:
http://xfs.org/index.php/XFS_FAQ#Q:_What_information_should_I_include_when_reporting_a_problem.3F
Thanks,
-Eric
> in regular intervals.
>
> Can you help me with this? Thanks in advance.
>
> Regards,
> Milan Holzäpfel
>
>
>
> [1]: http://oss.sgi.com/archives/xfs/2012-06/msg00362.html
> [2]: http://oss.sgi.com/archives/xfs/2012-09/msg00260.html
>
>
>
|