XFS lockups in 2.6.17.13

[Martin Orda <as_lkml@bgnet.pl>]

Morning,

Please do Cc me as I'm not subscribed. Thank you.

I've been experiencing issues with the XFS filesystem. I use SATA
Seagate ST3120026AS disk on which I have two partitions. First one is
used as XFS root partition and the second one is a LVM physical volume.

On my logical volumes I run XFS. I've started getting the lockups /
oopses after having been running the system for a couple of days. It
either results in the complete lock-up or IO operations become very
unstable and dmesg showing errors shown below.

Details:
distro:		debian/sid
kernel:		2.6.17.13
xfsprogs:	2.8.11-1
lvm2:		2.02.06-2

I've tried rebooting into single mode to minimize the number of daemons
running and then trying to create files on the partition affected (/usr)
reading them, listing contents, etc. Nothing happened. But it started to
show errors once I issued apt-get install something.

After that when I tried to ls /usr I got the IO Error message. Please
find the kernel output below.

I've gotten this error several times:

0x0: 00 00 00 00 01 06 6f 47 1d 46 63 41 74 6f 6d 69 
Filesystem "dm-0": XFS internal error xfs_da_do_buf(2) at line 2212 of file fs/xfs/xfs_da_btree.c.  Caller 0xc030193d
 <c0312b8a> xfs_corruption_error+0xe2/0x10a  <c030193d> xfs_da_read_buf+0x30/0x35
 <c033e474> kmem_zone_alloc+0x54/0xcb  <c0300ed4> xfs_da_buf_make+0xe2/0x12f
 <c0301887> xfs_da_do_buf+0x966/0x98d  <c030193d> xfs_da_read_buf+0x30/0x35
 <c030193d> xfs_da_read_buf+0x30/0x35  <c030c4f6> xfs_dir2_node_addname+0x480/0xb0e
 <c030c4f6> xfs_dir2_node_addname+0x480/0xb0e  <c0325743> xlog_write+0x70d/0x7f5
 <c0305bf5> xfs_dir2_isleaf+0x1b/0x50  <c0306553> xfs_dir2_createname+0xfc/0x131
 <c033e506> kmem_zone_zalloc+0x1b/0x43  <c033c268> xfs_link+0x31b/0x4fb
 <c03450f8> xfs_vn_link+0x43/0x90  <c01768cb> mntput_no_expire+0x13/0x72
 <c016b2ef> link_path_walk+0x65/0xcb  <c03380db> xfs_access+0x34/0x3a
 <c0345381> xfs_vn_permission+0x0/0x13  <c0345390> xfs_vn_permission+0xf/0x13
 <c0169105> permission+0x9f/0xa4  <c0169227> vfs_link+0x11d/0x168
 <c016bfa3> sys_linkat+0xdd/0xf9  <c015b8bd> sys_utime+0x48/0x12c
 <c041311d> _atomic_dec_and_lock+0x2d/0x50  <c01729f3> dput+0xfa/0x130
 <c015d122> __fput+0x131/0x194  <c016bfee> sys_link+0x2f/0x33
 <c0102af3> sysenter_past_esp+0x54/0x75


And this one just once:

xfs_da_do_buf: bno 16777216
dir: inode 17256410
Filesystem "dm-0": XFS internal error xfs_da_do_buf(1) at line 2119 of file fs/xfs/xfs_da_btree.c.  Caller 0xc030193d
 <c030123e> xfs_da_do_buf+0x31d/0x98d  <c030193d> xfs_da_read_buf+0x30/0x35
 <c030193d> xfs_da_read_buf+0x30/0x35  <c030d4e7> xfs_dir2_leafn_lookup_int+0x2c0/0x4f6
 <c030d4e7> xfs_dir2_leafn_lookup_int+0x2c0/0x4f6  <c030193d> xfs_da_read_buf+0x30/0x35
 <c030b88e> xfs_dir2_node_removename+0x30a/0x4ef  <c030b88e> xfs_dir2_node_removename+0x30a/0x4ef
 <c0306777> xfs_dir2_removename+0xee/0xf0  <c033be2e> xfs_remove+0x389/0x4a8
 <c0345381> xfs_vn_permission+0x0/0x13  <c034515c> xfs_vn_unlink+0x17/0x3b
 <c01768cb> mntput_no_expire+0x13/0x72  <c016b2ef> link_path_walk+0x65/0xcb
 <c03329b1> xfs_trans_unlocked_item+0x28/0x3c  <c03380db> xfs_access+0x34/0x3a
 <c0345381> xfs_vn_permission+0x0/0x13  <c0345390> xfs_vn_permission+0xf/0x13
 <c0169105> permission+0x9f/0xa4  <c01696d2> may_delete+0x2b/0x10d
 <c0169c8f> vfs_unlink+0x9c/0xf5  <c016ba7b> do_unlinkat+0xaf/0x12c
 <c0102af3> sysenter_past_esp+0x54/0x75 
Filesystem "dm-0": XFS internal error xfs_trans_cancel at line 1150 of file fs/xfs/xfs_trans.c.  Caller 0xc033be8b
 <c0331cea> xfs_trans_cancel+0xe1/0x102  <c033be8b> xfs_remove+0x3e6/0x4a8
 <c033be8b> xfs_remove+0x3e6/0x4a8  <c0345381> xfs_vn_permission+0x0/0x13
 <c034515c> xfs_vn_unlink+0x17/0x3b  <c01768cb> mntput_no_expire+0x13/0x72
 <c016b2ef> link_path_walk+0x65/0xcb  <c03329b1> xfs_trans_unlocked_item+0x28/0x3c
 <c03380db> xfs_access+0x34/0x3a  <c0345381> xfs_vn_permission+0x0/0x13
 <c0345390> xfs_vn_permission+0xf/0x13  <c0169105> permission+0x9f/0xa4
 <c01696d2> may_delete+0x2b/0x10d  <c0169c8f> vfs_unlink+0x9c/0xf5
 <c016ba7b> do_unlinkat+0xaf/0x12c  <c0102af3> sysenter_past_esp+0x54/0x75
xfs_force_shutdown(dm-0,0x8) called from line 1151 of file fs/xfs/xfs_trans.c.  Return address = 0xc0348379
Filesystem "dm-0": Corruption of in-memory data detected.  Shutting down filesystem: dm-0
Please umount the filesystem, and rectify the problem(s)

This problem affects 2 of my systems which don't run the same hardware
but have roughtly the same software and  config. Can you please advice
if this is related to a bug in kernel or userland and suggest some fix
for that if possible?

BTW: please let me know if I've forgotten to mention some details that
you could find helpful. Thanks for your help.

-- 
Kind regards,
Martin Orda
http://securityshells.com

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/