XFS internal error XFS_WANT_CORRUPTED_RETURN

DISCUSSION Markus Schoder 3 participants 5 messages Aug 23, 2007 – Aug 24, 2007

Share & tools

Markus Schoder

author

2007-08-23 20:33 UTC #0 lore raw

Got a bunch of the below errors in the log. It occured during a Debian upgrade with aptitude. xfs_repair found a lost inode.

Kernel is stock 2.6.22.4 plus CFS patch. This is a 64bit kernel on an amd64 processor.

I am glad to provide more information if required.

Would be nice to keep me CC'ed if this is not against policy.

--
Markus

Aug 23 00:15:09 gondolin kernel: [17152.830267] XFS internal error XFS_WANT_CORRUPTED_RETURN at line 281 of file fs/xfs/xfs_alloc.c.  Caller 0xffffffff802eb061
Aug 23 00:15:09 gondolin kernel: [17152.830274]
Aug 23 00:15:09 gondolin kernel: [17152.830275] Call Trace:
Aug 23 00:15:09 gondolin kernel: [17152.830295]  [<ffffffff802e9bf7>] xfs_alloc_fixup_trees+0x307/0x3a0
Aug 23 00:15:09 gondolin kernel: [17152.830301]  [<ffffffff80303ddd>] xfs_btree_setbuf+0x2d/0xb0
Aug 23 00:15:09 gondolin kernel: [17152.830306]  [<ffffffff802eb061>] xfs_alloc_ag_vextent_near+0x581/0x9b0
Aug 23 00:15:09 gondolin kernel: [17152.830313]  [<ffffffff802ebc95>] xfs_alloc_ag_vextent+0xd5/0x130
Aug 23 00:15:09 gondolin kernel: [17152.830316]  [<ffffffff802ec4f5>] xfs_alloc_vextent+0x275/0x460
Aug 23 00:15:09 gondolin kernel: [17152.830322]  [<ffffffff802f9630>] xfs_bmap_btalloc+0x410/0x7c0
Aug 23 00:15:09 gondolin kernel: [17152.830327]  [<ffffffff8032d6fa>] xfs_mod_incore_sb_batch+0xea/0x130
Aug 23 00:15:09 gondolin kernel: [17152.830335]  [<ffffffff802f6d1e>] xfs_bmap_isaeof+0x7e/0xd0
Aug 23 00:15:09 gondolin kernel: [17152.830344]  [<ffffffff802fd8c9>] xfs_bmapi+0xc89/0x1340
Aug 23 00:15:09 gondolin kernel: [17152.830359]  [<ffffffff80330298>] xfs_trans_reserve+0xa8/0x210
Aug 23 00:15:09 gondolin kernel: [17152.830364]  [<ffffffff80320e58>] xfs_iomap_write_direct+0x2e8/0x500
Aug 23 00:15:09 gondolin kernel: [17152.830374]  [<ffffffff8032028c>] xfs_iomap+0x31c/0x390
Aug 23 00:15:09 gondolin kernel: [17152.830383]  [<ffffffff8033dafa>] xfs_map_blocks+0x3a/0x80
Aug 23 00:15:09 gondolin kernel: [17152.830387]  [<ffffffff8033ec3e>] xfs_page_state_convert+0x2be/0x630
Aug 23 00:15:09 gondolin kernel: [17152.830394]  [<ffffffff802a23cc>] alloc_buffer_head+0x4c/0x80
Aug 23 00:15:09 gondolin kernel: [17152.830398]  [<ffffffff802a2be0>] alloc_page_buffers+0x60/0xe0
Aug 23 00:15:09 gondolin kernel: [17152.830403]  [<ffffffff8033f10f>] xfs_vm_writepage+0x6f/0x120
Aug 23 00:15:09 gondolin kernel: [17152.830408]  [<ffffffff8025dfba>] __writepage+0xa/0x30
Aug 23 00:15:09 gondolin kernel: [17152.830411]  [<ffffffff8025e5fe>] write_cache_pages+0x23e/0x330
Aug 23 00:15:09 gondolin kernel: [17152.830415]  [<ffffffff8025dfb0>] __writepage+0x0/0x30
Aug 23 00:15:09 gondolin kernel: [17152.830424]  [<ffffffff8025e740>] do_writepages+0x20/0x40
Aug 23 00:15:09 gondolin kernel: [17152.830427]  [<ffffffff80259715>] __filemap_fdatawrite_range+0x75/0xb0
Aug 23 00:15:09 gondolin kernel: [17152.830434]  [<ffffffff802a1975>] do_fsync+0x45/0xe0
Aug 23 00:15:09 gondolin kernel: [17152.830438]  [<ffffffff8026e2b6>] sys_msync+0x166/0x1d0
Aug 23 00:15:09 gondolin kernel: [17152.830444]  [<ffffffff8021e302>] ia32_sysret+0x0/0xa

David Chinner

2007-08-24 01:43 UTC #1 lore raw

Quoted 2 lines · Markus Schoder

Got a bunch of the below errors in the log. It occured during a Debian upgrade with aptitude. xfs_repair found a lost inode.

Can you post the output of xfs_repair if you still have it? It would be handy to correlæte the shutdown to an actual error on disk....

Quoted 2 lines

Kernel is stock 2.6.22.4 plus CFS patch. This is a 64bit kernel on an amd64 processor.

Hmmmm - so not exactly a stock kernel then. Which of the 19 versions the CFS patch was applied? Is it reproducable or only a one-off? if it's not a one-off problem, have you seen the problem without the CFS patch?

Cheers,

Dave.

Signature · 3 lines

Dave Chinner
Principal Engineer
SGI Australian Software Group

Markus Schoder

author

2007-08-24 19:19 UTC #2 lore raw

Quoted 7 lines · David Chinner

Quoted 2 lines · Markus Schoder
Got a bunch of the below errors in the log. It occured during a Debian upgrade with aptitude. xfs_repair found a lost inode.
Can you post the output of xfs_repair if you still have it? It would be handy to correlæte the shutdown to an actual error on disk....

Unfortunately I don't have the output anymore. From memory I recall there were two messsages indicating problems:

1. A message about rebuilding a directory at inode 128.
2. A lost inode that was moved to lost+found.

Other than that there were only the usual progress messages.

Quoted 7 lines

Quoted 2 lines
Kernel is stock 2.6.22.4 plus CFS patch. This is a 64bit kernel on an amd64 processor.
Hmmmm - so not exactly a stock kernel then. Which of the 19 versions the CFS patch was applied? Is it reproducable or only a one-off? if it's not a one-off problem, have you seen the problem without the CFS patch?

It is version 19.1 and the problem was a one-off never seen it before or since.

Signature · 1 line

Markus

Timothy Shimmin

2007-08-24 02:03 UTC #3 lore raw

In reply to Markus Schoder

Quoted 33 lines · Markus Schoder

Got a bunch of the below errors in the log. It occured during a Debian upgrade with aptitude. xfs_repair found a lost inode.

Kernel is stock 2.6.22.4 plus CFS patch. This is a 64bit kernel on an amd64 processor.

I am glad to provide more information if required.

Would be nice to keep me CC'ed if this is not against policy.

--
Markus

Aug 23 00:15:09 gondolin kernel: [17152.830267] XFS internal error XFS_WANT_CORRUPTED_RETURN at line 281 of file fs/xfs/xfs_alloc.c.  Caller 0xffffffff802eb061
Aug 23 00:15:09 gondolin kernel: [17152.830274]
Aug 23 00:15:09 gondolin kernel: [17152.830275] Call Trace:
Aug 23 00:15:09 gondolin kernel: [17152.830295]  [<ffffffff802e9bf7>] xfs_alloc_fixup_trees+0x307/0x3a0
Aug 23 00:15:09 gondolin kernel: [17152.830301]  [<ffffffff80303ddd>] xfs_btree_setbuf+0x2d/0xb0
Aug 23 00:15:09 gondolin kernel: [17152.830306]  [<ffffffff802eb061>] xfs_alloc_ag_vextent_near+0x581/0x9b0
Aug 23 00:15:09 gondolin kernel: [17152.830313]  [<ffffffff802ebc95>] xfs_alloc_ag_vextent+0xd5/0x130
Aug 23 00:15:09 gondolin kernel: [17152.830316]  [<ffffffff802ec4f5>] xfs_alloc_vextent+0x275/0x460
Aug 23 00:15:09 gondolin kernel: [17152.830322]  [<ffffffff802f9630>] xfs_bmap_btalloc+0x410/0x7c0
Aug 23 00:15:09 gondolin kernel: [17152.830327]  [<ffffffff8032d6fa>] xfs_mod_incore_sb_batch+0xea/0x130
Aug 23 00:15:09 gondolin kernel: [17152.830335]  [<ffffffff802f6d1e>] xfs_bmap_isaeof+0x7e/0xd0
Aug 23 00:15:09 gondolin kernel: [17152.830344]  [<ffffffff802fd8c9>] xfs_bmapi+0xc89/0x1340
Aug 23 00:15:09 gondolin kernel: [17152.830359]  [<ffffffff80330298>] xfs_trans_reserve+0xa8/0x210
Aug 23 00:15:09 gondolin kernel: [17152.830364]  [<ffffffff80320e58>] xfs_iomap_write_direct+0x2e8/0x500
Aug 23 00:15:09 gondolin kernel: [17152.830374]  [<ffffffff8032028c>] xfs_iomap+0x31c/0x390
Aug 23 00:15:09 gondolin kernel: [17152.830383]  [<ffffffff8033dafa>] xfs_map_blocks+0x3a/0x80
Aug 23 00:15:09 gondolin kernel: [17152.830387]  [<ffffffff8033ec3e>] xfs_page_state_convert+0x2be/0x630
Aug 23 00:15:09 gondolin kernel: [17152.830394]  [<ffffffff802a23cc>] alloc_buffer_head+0x4c/0x80
Aug 23 00:15:09 gondolin kernel: [17152.830398]  [<ffffffff802a2be0>] alloc_page_buffers+0x60/0xe0
Aug 23 00:15:09 gondolin kernel: [17152.830403]  [<ffffffff8033f10f>] xfs_vm_writepage+0x6f/0x120

Quoted 9 lines

Aug 23 00:15:09 gondolin kernel: [17152.830408]  [<ffffffff8025dfba>] __writepage+0xa/0x30
Aug 23 00:15:09 gondolin kernel: [17152.830411]  [<ffffffff8025e5fe>] write_cache_pages+0x23e/0x330
Aug 23 00:15:09 gondolin kernel: [17152.830415]  [<ffffffff8025dfb0>] __writepage+0x0/0x30
Aug 23 00:15:09 gondolin kernel: [17152.830424]  [<ffffffff8025e740>] do_writepages+0x20/0x40
Aug 23 00:15:09 gondolin kernel: [17152.830427]  [<ffffffff80259715>] __filemap_fdatawrite_range+0x75/0xb0
Aug 23 00:15:09 gondolin kernel: [17152.830434]  [<ffffffff802a1975>] do_fsync+0x45/0xe0
Aug 23 00:15:09 gondolin kernel: [17152.830438]  [<ffffffff8026e2b6>] sys_msync+0x166/0x1d0
Aug 23 00:15:09 gondolin kernel: [17152.830444]  [<ffffffff8021e302>] ia32_sysret+0x0/0xa

It looks a lot like a reported bug:
Suse#198124 and sgi-pv#956334

Where we had corruption in freespace btrees.

Can you also run xfs_check before running xfs_repair.

--Tim

Markus Schoder

author

2007-08-24 19:23 UTC #4 lore raw

Quoted 6 lines · Timothy Shimmin

It looks a lot like a reported bug:
   Suse#198124 and sgi-pv#956334
Where we had corruption in freespace btrees.
Can you also run xfs_check before running xfs_repair.

I actually tried to run xfs_check before xfs_repair but it seg-faulted.

--
Markus