XFS trouble on 2.6.0-test11
Congratulations on your 2.6.0 release. I hate to spoil it but I'm having some serious XFS trouble. I was attempting to build an rpm of the latest kernel, when XFS crashed,
and over the next few boots ( while I was trying to fix program after program) completely managed to mess up my system to the point when it no longer boots. In fact, it doesn't even mount most of the time. Anyway, I managed to save the original error, so here it is (see below). Any ideas as to what's wrong and how I can fix my filesystem? It's a miracle it actually mounted under Knoppix so I could retrieve the error - it failed the last few times I tried to mount. Apparently the recovery during mount is not doing its job, since the filesystem continually corrupts everything I touch and is causing lots of damage. Note: I have not had any trouble with XFS before and I have been running the 2.6.0-test11 kernel for a while (1-2 weeks?)
I will not be able to reply to this message for at least 24 hrs as I have to catch a flight very soon. Unfortunately my computer's not coming along, so I won't be able to help with any testing for a month.
Here's the error I got:
Filesystem "hda7": XFS internal error xfs_alloc_read_agf at line 2208 of file fs/xfs/xfs_alloc.c. Caller 0xc01bb7ea
Call Trace: [<c01bbbc2>] xfs_alloc_read_agf+0xe2/0x1e0 [<c01bb7ea>] xfs_alloc_fix_freelist+0x45a/0x470 [<c01bb7ea>] xfs_alloc_fix_freelist+0x45a/0x470 [<c01bb7ea>] xfs_alloc_fix_freelist+0x45a/0x470 [<c01ff87b>] xlog_grant_log_space+0x12b/0x380 [<c0142a5c>] cache_grow+0x17c/0x290 [<c01bc0e3>] xfs_free_extent+0x93/0xf0 [<c01ed116>] xfs_efd_init+0x116/0x140 [<c020e518>] xfs_trans_get_efd+0x38/0x50 [<c0205348>] xlog_recover_process_efi+0x188/0x200 [<c0205415>] xlog_recover_process_efis+0x55/0xa0 [<c0206b74>] xlog_recover_finish+0x24/0xf0 [<c01b0ff3>] xfs_qm_newmount+0x93/0x190 [<c01fd40c>] xfs_log_mount_finish+0x2c/0x30 [<c020839f>] xfs_mountfs+0x8cf/0xfd0 [<c0225710>] xfs_setsize_buftarg+0x40/0x80 [<c01fa8fe>] xfs_ioinit+0x1e/0x40 [<c02102df>] xfs_mount+0x33f/0x5e0 [<c02263b4>] vfs_mount+0x34/0x40 [<c02263b4>] vfs_mount+0x34/0x40 [<c022619b>] linvfs_fill_super+0x9b/0x220 [<c02373b7>] snprintf+0x27/0x30 [<c018b262>] disk_name+0x62/0xb0 [<c015cfb5>] sb_set_blocksize+0x25/0x60 [<c015c954>] get_sb_bdev+0x124/0x160 [<c022634f>] linvfs_get_sb+0x2f/0x60 [<c0226100>] linvfs_fill_super+0x0/0x220 [<c015cbc3>] do_kern_mount+0x63/0x110 [<c017290a>] do_add_mount+0x6a/0x150 [<c0172c30>] do_mount+0x150/0x1a0 [<c0172a70>] copy_mount_options+0x80/0xf0 [<c0172fff>] sys_mount+0xbf/0x140 [<c0420c1f>] do_mount_root+0x2f/0xa0 [<c0420ce4>] mount_block_root+0x54/0x120 [<c0420f26>] mount_root+0x36/0x40 [<c0420f75>] prepare_namespace+0x45/0x150 [<c01050d7>] init+0x37/0x160 [<c01050a0>] init+0x0/0x160 [<c0109289>] kernel_thread_helper+0x5/0xc
Failed mount attempts also talk about error in xfs_alloc_read_agf.
Yes unfortunately this has been cropping in a few places recently, including 2.4.
We don't have much info as to what is going wrong at the point other than the panic is due to bad data on disk. Tracking down when the bad data happened it always difficult.
If anybody else hits this attach the output of xfs_check to:
http://oss.sgi.com/bugzilla/show_bug.cgi?id=296
If you're system is accessible and you would be willing to let somebody poke around you're system with xfs_db try to find us on irc: FreeNode #xfs
Quoted 81 lines · Knoppix User
Congratulations on your 2.6.0 release. I hate to spoil it but I'm having some serious XFS trouble.
I was attempting to build an rpm of the latest kernel, when XFS crashed, and over the next few boots ( while I was trying to fix program after program) completely managed
to mess up my system to the point when it no longer boots. In fact, it doesn't even mount
most of the time. Anyway, I managed to save the original error, so here it is (see below).
Any ideas as to what's wrong and how I can fix my filesystem? It's a miracle it actually mounted
under Knoppix so I could retrieve the error - it failed the last few times I tried to mount.
Apparently the recovery during mount is not doing its job, since the filesystem continually
corrupts everything I touch and is causing lots of damage. Note: I have not had any trouble
with XFS before and I have been running the 2.6.0-test11 kernel for a while (1-2 weeks?)I will not be able to reply to this message for at least 24 hrs as I have to catch a flight very soon.
Unfortunately my computer's not coming along, so I won't be able to help with any testing for a month.Here's the error I got:
Filesystem "hda7": XFS internal error
xfs_alloc_read_agfat line 2208 of filefs/xfs/xfs_alloc.c. Caller0xc01bb7eaCall Trace: [<c01bbbc2>] xfs_alloc_read_agf+0xe2/0x1e0 [<c01bb7ea>] xfs_alloc_fix_freelist+0x45a/0x470 [<c01bb7ea>] xfs_alloc_fix_freelist+0x45a/0x470 [<c01bb7ea>] xfs_alloc_fix_freelist+0x45a/0x470 [<c01ff87b>] xlog_grant_log_space+0x12b/0x380 [<c0142a5c>] cache_grow+0x17c/0x290 [<c01bc0e3>] xfs_free_extent+0x93/0xf0 [<c01ed116>] xfs_efd_init+0x116/0x140 [<c020e518>] xfs_trans_get_efd+0x38/0x50 [<c0205348>] xlog_recover_process_efi+0x188/0x200 [<c0205415>] xlog_recover_process_efis+0x55/0xa0 [<c0206b74>] xlog_recover_finish+0x24/0xf0 [<c01b0ff3>] xfs_qm_newmount+0x93/0x190 [<c01fd40c>] xfs_log_mount_finish+0x2c/0x30 [<c020839f>] xfs_mountfs+0x8cf/0xfd0 [<c0225710>] xfs_setsize_buftarg+0x40/0x80 [<c01fa8fe>] xfs_ioinit+0x1e/0x40 [<c02102df>] xfs_mount+0x33f/0x5e0 [<c02263b4>] vfs_mount+0x34/0x40 [<c02263b4>] vfs_mount+0x34/0x40 [<c022619b>] linvfs_fill_super+0x9b/0x220 [<c02373b7>] snprintf+0x27/0x30 [<c018b262>] disk_name+0x62/0xb0 [<c015cfb5>] sb_set_blocksize+0x25/0x60 [<c015c954>] get_sb_bdev+0x124/0x160 [<c022634f>] linvfs_get_sb+0x2f/0x60 [<c0226100>] linvfs_fill_super+0x0/0x220 [<c015cbc3>] do_kern_mount+0x63/0x110 [<c017290a>] do_add_mount+0x6a/0x150 [<c0172c30>] do_mount+0x150/0x1a0 [<c0172a70>] copy_mount_options+0x80/0xf0 [<c0172fff>] sys_mount+0xbf/0x140 [<c0420c1f>] do_mount_root+0x2f/0xa0 [<c0420ce4>] mount_block_root+0x54/0x120 [<c0420f26>] mount_root+0x36/0x40 [<c0420f75>] prepare_namespace+0x45/0x150 [<c01050d7>] init+0x37/0x160 [<c01050a0>] init+0x0/0x160 [<c0109289>] kernel_thread_helper+0x5/0xcFailed mount attempts also talk about error in
xfs_alloc_read_agf.-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/