[mythtv-users] OT: Annoying XFS corruptions on mythtv box?
John Drescher
drescherjm at gmail.com
Mon Mar 9 13:57:18 UTC 2009
On Mon, Mar 9, 2009 at 9:20 AM, Mike Dent <mcdent at gmail.com> wrote:
> Hi,
> not sure where to go with this one. I have a mythtv system 0.21 built
> on hardware about 18 months old.
> This is the kernel I am running:
> Linux version 2.6.24-21-generic (buildd at palmer) (gcc version 4.2.3
> (Ubuntu 4.2.3-2ubuntu7)) #1 SMP Mon Aug 25 17:32:09 UTC 2008 (Ubuntu
> 2.6.24-21.42-generic)
> and the processor is a "model name : AMD Athlon(tm) 64 Processor 3200+"
>
> The disk with the corruption is 250gb ata3.00: ATA-6: ST3200822AS,
> 3.01, max UDMA/133 and it seems to be the XFS partition which is
> being corrupted /dev/sda3 which is mounted on /var/lib
>
>
> Here is the corruption I see in the log
>
> [350535.066389] XFS internal error XFS_WANT_CORRUPTED_RETURN at line
> 281 of file /build/buildd/linux-2.6.24/fs/xfs/xfs_alloc.c. Caller
> 0xf8f96168
> [350535.066403] Pid: 18834, comm: mythbackend Tainted: P
> 2.6.24-21-generic #1
> [350535.066434] [<f8f945d6>] xfs_alloc_fixup_trees+0x2e6/0x3a0 [xfs]
> [350535.066491] [<f8f96168>] xfs_alloc_ag_vextent_near+0x448/0xa10 [xfs]
> [350535.066566] [<f8f96168>] xfs_alloc_ag_vextent_near+0x448/0xa10 [xfs]
> [350535.066715] [<f8f967f7>] xfs_alloc_ag_vextent+0xc7/0x120 [xfs]
> [350535.066772] [<f8f970c5>] xfs_alloc_vextent+0x275/0x500 [xfs]
> [350535.066879] [<f8faae93>] xfs_bmap_btalloc+0x513/0xc40 [xfs]
> [350535.066945] [<f8fc90bd>] xfs_iext_bno_to_ext+0xad/0x1f0 [xfs]
> [350535.067163] [<f8fabd64>] xfs_bmapi+0x774/0x1710 [xfs]
> [350535.067229] [<f8fc90bd>] xfs_iext_bno_to_ext+0xad/0x1f0 [xfs]
> [350535.067368] [<f8fa210f>] xfs_bmap_search_extents+0x6f/0x150 [xfs]
> [350535.067605] [<f8fd03c9>] xfs_iomap_write_allocate+0x2d9/0x500 [xfs]
> [350535.067856] [<f8fcf4b6>] xfs_iomap+0x4b6/0x510 [xfs]
> [350535.067951] [<c0122cb4>] find_busiest_group+0x1c4/0x780
> [350535.068039] [<f8feef63>] xfs_map_blocks+0x43/0x90 [xfs]
> [350535.068121] [<f8ff0821>] xfs_page_state_convert+0x401/0x820 [xfs]
> [350535.068186] [<c010336e>] __switch_to+0x9e/0x150
> [350535.068334] [<f8ff0d71>] xfs_vm_writepage+0x61/0xf0 [xfs]
> [350535.068414] [<c0173aa8>] __writepage+0x8/0x30
> [350535.068424] [<c01740af>] write_cache_pages+0x21f/0x310
> [350535.068438] [<c0173aa0>] __writepage+0x0/0x30
> [350535.068557] [<c01741c0>] generic_writepages+0x20/0x30
> [350535.068571] [<c01741fb>] do_writepages+0x2b/0x50
> [350535.068585] [<c016e44a>] __filemap_fdatawrite_range+0x7a/0xa0
> [350535.068659] [<c016e723>] filemap_fdatawrite+0x23/0x30
> [350535.068678] [<c01b1f7e>] do_fsync+0x4e/0xb0
> [350535.068704] [<c01b2005>] __do_fsync+0x25/0x40
> [350535.068717] [<c01043b2>] sysenter_past_esp+0x6b/0xa9
> [350535.068794] =======================
>
> And I see
>
> [1846802.621851] xfs_force_shutdown(sda3,0x8) called from line 1164 of
> file /build/buildd/linux-2.6.24/fs/xfs/xfs_trans.c. Return address =
> 0xf8fe2429
> [1846802.651842] Filesystem "sda3": Corruption of in-memory data
> detected. Shutting down filesystem: sda3
> [1846802.651854] Please umount the filesystem, and rectify the problem(s)
>
>
> Has anybody else seen anything like this?
>
I have seen this in some older kernels. I would upgrade to at minimum
2.6.26 at this time.
John
More information about the mythtv-users
mailing list