Details
-
Bug
-
Resolution: Fixed
-
Major
-
Lustre 2.5.3
-
3
-
16196
Description
During Tests, we hit something like following:
<2>LDISKFS-fs error (device dm-2): __ldiskfs_ext_check_block: bad header/extent in inode #659: invalid magic - magic e000, entries 456, max 0(0), depth 51424(0)
<3>Aborting journal on device dm-2-8.
<2>LDISKFS-fs error (device dm-2) in ldiskfs_free_blocks: Journal has aborted
<2>LDISKFS-fs error (device dm-2) in ldiskfs_free_blocks: Journal has aborted
<2>LDISKFS-fs (dm-2): Remounting filesystem read-only
<2>LDISKFS-fs error (device dm-2) in ldiskfs_free_blocks: Journal has aborted
<2>LDISKFS-fs error (device dm-2) in ldiskfs_ext_remove_space: Journal has aborted
<2>LDISKFS-fs error (device dm-2) in ldiskfs_reserve_inode_write: Journal has aborted
<2>LDISKFS-fs error (device dm-2) in ldiskfs_ext_truncate: Journal has aborted
<4>LDISKFS-fs warning (device dm-2): ldiskfs_delete_inode: couldn't extend journal (err -5)
<3>LDISKFS-fs (dm-2): Inode 280 (ffff8803a9ecb6d8): orphan list check failed!
Some bad thing happen which forces filesystem to readonly, and there are still in memory orphan inode that was not cleared, which cause following problem:
<4>Pid: 45622, comm: umount Not tainted 2.6.32-431.17.1.el6_lustre.2.5.18.ddn2.x86_64 #1 Dell Inc. PowerEdge R620/01W23F
<4>RIP: 0010:[<ffffffffa169081a>] [<ffffffffa169081a>] ldiskfs_put_super+0x33a/0x380 [ldiskfs]
<4>RSP: 0018:ffff88027ecf39f8 EFLAGS: 00010296
<4>RAX: 003fffffffffffd4 RBX: ffff88102359c800 RCX: 00400000000000a4
<4>RDX: 0000000000000000 RSI: 0000000000000046 RDI: ffffffffa16a55b8
<4>RBP: ffff88027ecf3a38 R08: 0000000000000000 R09: ffffffff81645da0
<4>R10: 0000000000000001 R11: 0000000000000000 R12: ffff88102359c000
<4>R13: ffff88102359c980 R14: ffff88102359c9f0 R15: 004000000000006c
<4>FS: 00007fe81d661740(0000) GS:ffff880061c00000(0000) knlGS:0000000000000000
<4>CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
<4>CR2: 00007fff13c76010 CR3: 0000001d707f0000 CR4: 00000000001407f0
<4>DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
<4>DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
<4>Process umount (pid: 45622, threadinfo ffff88027ecf2000, task ffff8810258ac040)
<4>Stack:
<4> ffff880200000000 ffff88027ecf39f8 ffff88102359c000 ffff88102359c000
<4><d> ffffffffa169d5a0 ffffffff81c06500 ffff88102359c000 ffff880f0371e138
<4><d> ffff88027ecf3a58 ffffffff8118af0b ffff881fe54cb540 0000000000000003
<4>Call Trace:
<4> [<ffffffff8118af0b>] generic_shutdown_super+0x5b/0xe0
<4> [<ffffffff8118afc1>] kill_block_super+0x31/0x50
<4> [<ffffffff8118b797>] deactivate_super+0x57/0x80
<4> [<ffffffff811aa79f>] mntput_no_expire+0xbf/0x110
<4> [<ffffffffa1719d99>] osd_umount+0x79/0x150 [osd_ldiskfs]
<4> [<ffffffffa171e9b7>] osd_device_fini+0x147/0x190 [osd_ldiskfs]
<4> [<ffffffffa103b973>] class_cleanup+0x573/0xd30 [obdclass]
<4> [<ffffffffa100e366>] ? class_name2dev+0x56/0xe0 [obdclass]
<4> [<ffffffffa103d69a>] class_process_config+0x156a/0x1ad0 [obdclass]
<4> [<ffffffffa1035ff3>] ? lustre_cfg_new+0x2d3/0x6e0 [obdclass]
<4> [<ffffffffa103dd79>] class_manual_cleanup+0x179/0x6f0 [obdclass]
<4> [<ffffffffa100c83b>] ? class_export_put+0x10b/0x2c0 [obdclass]
<4> [<ffffffffa1723c65>] osd_obd_disconnect+0x1c5/0x1d0 [osd_ldiskfs]
<4> [<ffffffffa104031b>] lustre_put_lsi+0x1ab/0x11a0 [obdclass]
<4> [<ffffffffa10488c8>] lustre_common_put_super+0x5d8/0xbf0 [obdclass]
<4> [<ffffffffa1070f6d>] server_put_super+0x1bd/0xf60 [obdclass]
<4> [<ffffffff8118af0b>] generic_shutdown_super+0x5b/0xe0
<4> [<ffffffff8118aff6>] kill_anon_super+0x16/0x60
<4> [<ffffffffa103fc26>] lustre_kill_super+0x36/0x60 [obdclass]
<4> [<ffffffff8118b797>] deactivate_super+0x57/0x80
<4> [<ffffffff811aa79f>] mntput_no_expire+0xbf/0x110
<4> [<ffffffff811ab2eb>] sys_umount+0x7b/0x3a0
<4> [<ffffffff8108a281>] ? sigprocmask+0x71/0x110
<4> [<ffffffff8100b072>] system_call_fastpath+0x16/0x1b
<4>Code: 01 00 00 4d 39 fe 75 11 4c 3b b3 f0 01 00 00 0f 84 81 fe ff ff 0f 0b eb fe 49 8d 87 68 ff ff ff 49 8d 4f 38 48 c7 c7 b8 55 6a a1 <48> 8b b0 d8 01 00 00 44 8b 88 1c 01 00 00 44 0f b7 80 7e 01 00
<1>RIP [<ffffffffa169081a>] ldiskfs_put_super+0x33a/0x380 [ldiskfs]
<4> RSP <ffff88027ecf39f8>
This maybe because a free after accessing problem, from codes inode memory is freed, and in ext4_put_super, we will access it which maybe cause problem(I am not sure about this part analysis.)
But even above problem is not true, we still can run into:
J_ASSERT(list_empty(&sbi->s_orphan))
which will crash kernel, so we need fix this problem.