http://oss.sgi.com/bugzilla/show_bug.cgi?id=775
Summary: XFS internal errors when working with many small files
Product: Linux XFS
Version: Current
Platform: PC
OS/Version: Linux
Status: NEW
Severity: critical
Priority: P1
Component: XFS kernel code
AssignedTo: xfs-master@xxxxxxxxxxx
ReportedBy: igor@xxxxxxxx
XFS crashes during rsync process with a lot of files (linux installation) from
another filesystem to XFS.
There are a lot of these errors:
-------------------------------------------------------------------
Filesystem "sda": XFS internal error xfs_da_do_buf(2) at line 2085 of file
fs/xfs/xfs_da_btree.c. Caller 0xffffffff80383d8f
Call Trace:
[<ffffffff80383cd3>] xfs_da_do_buf+0x6c8/0x717
[<ffffffff80383d8f>] xfs_da_read_buf+0x24/0x29
[<ffffffff803ad40e>] xfs_trans_read_buf+0x47/0x319
[<ffffffff80383d8f>] xfs_da_read_buf+0x24/0x29
[<ffffffff80386bc2>] xfs_dir2_block_lookup_int+0x46/0x1ab
[<ffffffff80386bc2>] xfs_dir2_block_lookup_int+0x46/0x1ab
[<ffffffff80286118>] do_select+0x51e/0x54e
[<ffffffff80387568>] xfs_dir2_block_lookup+0x15/0x90
[<ffffffff80386613>] xfs_dir_lookup+0xca/0x129
[<ffffffff802292a9>] update_curr_load+0x71/0x87
[<ffffffff80229bbe>] __check_preempt_curr_fair+0x50/0x79
[<ffffffff803ae3f5>] xfs_dir_lookup_int+0x32/0xcf
[<ffffffff803b19ff>] xfs_lookup+0x5a/0x86
[<ffffffff803bf10c>] xfs_vn_lookup+0x36/0x72
[<ffffffff80280537>] do_lookup+0x18b/0x1de
[<ffffffff80281ee8>] __link_path_walk+0x30e/0xecc
[<ffffffff805e152e>] unix_stream_sendmsg+0x286/0x395
[<ffffffff80282b01>] link_path_walk+0x5b/0xef
[<ffffffff80282de1>] do_path_lookup+0x87/0x1dd
[<ffffffff802816f9>] getname+0xc9/0x1e5
[<ffffffff802837f4>] __user_walk_fd+0x37/0x59
[<ffffffff8027be23>] vfs_lstat_fd+0x18/0x47
[<ffffffff8027c021>] sys_newlstat+0x19/0x31
[<ffffffff802793c2>] vfs_write+0x12e/0x172
[<ffffffff80279913>] sys_write+0x45/0x73
[<ffffffff8020b7be>] system_call+0x7e/0x83
0x0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
---------------------------------------------------------------
As a result there are a lot of corrupted files that can't be removed in usual
way, although XFS itself didn't shutdown itself as it would do in case of other
errors...
xfs_repair did repaired it, showing a lot of `bad magic number on inode N`.
Well, it seems quite strange to me that XFS bug occurs again (see bug 773 for
another XFS problem on this server) that i'm starting to blame the hardware
backend, although RAID monitoring shows no errors(and I/O errors on device
aren't reported) and i haven't noticed any problems until i tried to backup my
root FS to XFS.
XFS partition is occupied to ~50% (1.7TB/3.5Tb) with big files (>15Gb) and i
haven't had absolutely no problems with them.
Give me a hint please :) If it's not XFS problem, i'll start searching in
another direction.
Hardware - see bug 773
FS created by:
mkfs.xfs -f -d agcount=900,su=4096,sw=28672 -l size=128m,su=128k /dev/sda
--
Configure bugmail: http://oss.sgi.com/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
|