scjody@clusterfs.com
2007-Jun-15 11:38 UTC
[Lustre-devel] [Bug 12746] New: kernel BUG at /root/cfs/b1_6/ldiskfs/ldiskfs/mballoc.c:2913 during sanity-quota
Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
https://bugzilla.lustre.org/show_bug.cgi?id=12746
In b1_6:
This only occurs during sanity-quota, so it could be a quota issue, but it
really looks like an mballoc issue:
Lustre: 5155:0:(debug.c:433:libcfs_debug_mark_buffer()) DEBUG MARKER: == test 0:
Set quota ============================== 13:37:02 (1181929022)
Lustre: 5168:0:(debug.c:433:libcfs_debug_mark_buffer()) DEBUG MARKER: == test 1:
Block hard limit (normal use and out of quota) ==== 13:37:03 (1181929023)
------------[ cut here ]------------
kernel BUG at /root/cfs/b1_6/ldiskfs/ldiskfs/mballoc.c:2913!
invalid operand: 0000 [#1]
SMP
Modules linked in: loop(U) mgs(U) mgc(U) lustre(U) obdfilter(U) ost(U)
fsfilt_ldiskfs(U) ldiskfs(U) mds(U) lov(U) osc(U) mdc(U) lquota(U) ptlrpc(U)
obdclass(U) lvfs(U) ksocklnd(U) lnet(U) libcfs(U) nfsd(U) exportfs(U) nfs(U)
lockd(U) nfs_acl(U) sunrpc(U) dm_mirror(U) dm_mod(U) button(U) battery(U) ac(U)
uhci_hcd(U) snd_ens1371(U) snd_rawmidi(U) snd_seq_device(U) snd_pcm_oss(U)
snd_mixer_oss(U) snd_pcm(U) snd_timer(U) snd_page_alloc(U) snd_ac97_codec(U)
snd(U) soundcore(U) pcnet32(U) mii(U) floppy(U) ext3(U) jbd(U) mptscsih(U)
mptsas(U) mptspi(U) mptfc(U) mptscsi(U) mptbase(U) sd_mod(U) scsi_mod(U)
CPU: 0
EIP: 0060:[<f1046078>] Tainted: GF VLI
EFLAGS: 00010202 (2.6.9-prep.qp2.2.5.11.3qsnet)
EIP is at ldiskfs_mb_mark_diskspace_used+0xa8/0x410 [ldiskfs]
eax: e58b9860 ebx: f1059d60 ecx: e78f1a00 edx: 00000001
esi: e8cf4200 edi: 00000000 ebp: 00000000 esp: e58b9810
ds: 007b es: 007b ss: 0068
Process ll_ost_io_02 (pid: 4950, threadinfo=e58b8000 task=e67751b0)
Stack: 00000000 00000003 00001112 00000000 e58b9860 e553edd4 00003111 f1046cdc
e78f1a00 e79d0000 e590d400 00000000 eb4732cc f1046ec4 e58b98f0 e78f1a00
eb4732cc e58b997c f10483b3 00000400 e8cf4200 e78f1a00 00001111 00003111
Call Trace:
[<f1046cdc>] ldiskfs_mb_use_inode_pa+0x4c/0xb0 [ldiskfs]
[<f1046ec4>] ldiskfs_mb_use_preallocated+0x134/0x140 [ldiskfs]
[<f10483b3>] ldiskfs_mb_new_blocks+0xb3/0x150 [ldiskfs]
[<f0e63082>] new_blocks+0xc2/0xe0 [fsfilt_ldiskfs]
[<f0e6329f>] ldiskfs_ext_new_extent_cb+0x1ff/0x870 [fsfilt_ldiskfs]
[<c02d1fb2>] __cond_resched+0x14/0x39
[<f103e67f>] ldiskfs_ext_find_extent+0x3bf/0x470 [ldiskfs]
[<f09a5611>] libcfs_debug_vmsg2+0x3f1/0x630 [libcfs]
[<f103fedf>] ldiskfs_ext_walk_space+0xaf/0x290 [ldiskfs]
[<f0e639e8>] fsfilt_map_nblocks+0xd8/0x1e0 [fsfilt_ldiskfs]
[<f0e630a0>] ldiskfs_ext_new_extent_cb+0x0/0x870 [fsfilt_ldiskfs]
[<f0e63bef>] fsfilt_ldiskfs_map_ext_inode_pages+0xff/0x350
[fsfilt_ldiskfs]
[<f0ea7ec6>] filter_direct_io+0x1e6/0x14e0 [obdfilter]
[<f0e60bd9>] fsfilt_ldiskfs_brw_start+0x129/0x510 [fsfilt_ldiskfs]
[<f0eaa23f>] filter_commitrw_write+0xfef/0x4157 [obdfilter]
[<c02d1fb2>] __cond_resched+0x14/0x39
[<f0e0f107>] ost_brw_write+0x24b7/0x3780 [ost]
[<c02d1fb2>] __cond_resched+0x14/0x39
[<f0c9249e>] lnet_prep_send+0x9e/0xa0 [lnet]
[<f0c9814d>] LNetPut+0x3fd/0xce0 [lnet]
[<f0e08120>] ost_bulk_timeout+0x0/0x200 [ost]
[<f0e08120>] ost_bulk_timeout+0x0/0x200 [ost]
[<f0e16664>] ost_handle+0x22b4/0x5410 [ost]
[<c01c188d>] vsnprintf+0x448/0x488
[<f0bb9ff9>] ptlrpc_server_handle_request+0x909/0x1c00 [ptlrpc]
[<c0128e69>] __mod_timer+0x101/0x10b
[<c011d758>] default_wake_function+0x0/0xc
[<f099c3a3>] cfs_block_sigs+0x73/0xa0 [libcfs]
[<f0bbc539>] ptlrpc_main+0x799/0xe80 [ptlrpc]
[<c011d758>] default_wake_function+0x0/0xc
[<f0bbbd90>] ptlrpc_retry_rqbds+0x0/0x10 [ptlrpc]
[<c02d3922>] ret_from_fork+0x6/0x14
[<f0bbbd90>] ptlrpc_retry_rqbds+0x0/0x10 [ptlrpc]
[<f0bbbda0>] ptlrpc_main+0x0/0xe80 [ptlrpc]
[<c01041f5>] kernel_thread_helper+0x5/0xb
Code: 02 00 00 8b 5b 28 89 d1 89 c2 c7 04 24 00 00 00 00 89 f0 ff 53 08 48 0f 85
74 02 00 00 4f 85 ff 75 b1 8b 44 24 10 39 78 34 74 08 <0f> 0b 61 0b 58 d5
04 f1
85 ff bb 86 ff ff ff 0f 84 23 02 00 00
<0>Fatal exception: panic in 5 seconds
Test output:
== test 1: Block hard limit (normal use and out of quota) ==== 13:37:03
(1181929023)
User quota (limit: 20000 kbytes)
Disk quotas for user quota_usr (uid 60000):
Filesystem blocks quota limit grace files quota limit grace
/mnt/lustre 0 0 20000 0 0 0
lustre-MDT0000_UUID
0 0 1000 0 0 0
lustre-OST0000_UUID
0 0 1000
lustre-OST0001_UUID
0 0 1000
lustre-OST0002_UUID
0 0 1000
Write ...
running as UID 60000, GID 60000
[dd] [if=/dev/zero] [of=/mnt/lustre/quota_test_dir/quota_tst10] [bs=1024]
[count=10000]
10000+0 records in
10000+0 records out
Done
Write out of block quota ...
running as UID 60000, GID 60000
[dd] [if=/dev/zero] [of=/mnt/lustre/quota_test_dir/quota_tst10] [bs=1024]
[count=10000] [seek=10000]