[LU-6192] seq_client_alloc_fid()) ASSERTION( seq != ((void *)0) ) failed

• Status: Resolved
• Project: Lustre
• Affects Version/s: Lustre 2.7.0
• Type: Bug
• Priority: Critical
• Reporter: Oleg Drokin
• Assignee: WC Triage
• Resolution: Duplicate
• Labels: zfs
• Severity: 3
• Bugzilla ID: 6,154

Description

Trying to see how sanity works on ZFS (and I also have DNE on - two MDS) trying to run simple sanity.sh in slow mode, test 31p causes MDS to crash 100% of the time with that assertion:

<4>[ 4377.551909] Lustre: DEBUG MARKER: == sanity test 31p: remove of open striped directory == 22:00:43 (1422846043)
<0>[ 4377.641621] LustreError: 14253:0:(fid_request.c:329:seq_client_alloc_fid()) ASSERTION( seq != ((void *)0) ) failed: 
<0>[ 4377.641922] LustreError: 14253:0:(fid_request.c:329:seq_client_alloc_fid()) LBUG
<4>[ 4377.642333] Pid: 14253, comm: mdt01_009
<4>[ 4377.642459] 
<4>[ 4377.642459] Call Trace:
<4>[ 4377.642681]  [<ffffffffa05e88a5>] libcfs_debug_dumpstack+0x55/0x80 [libcfs]
<4>[ 4377.642839]  [<ffffffffa05e8ea7>] lbug_with_loc+0x47/0xb0 [libcfs]
<4>[ 4377.642996]  [<ffffffffa0bb005c>] seq_client_alloc_fid+0x46c/0x470 [fid]
<4>[ 4377.643155]  [<ffffffff81175e1b>] ? __kmalloc+0x1bb/0x2a0
<4>[ 4377.643298]  [<ffffffffa12395dc>] osd_fid_alloc+0xbc/0xc0 [osd_zfs]
<4>[ 4377.643453]  [<ffffffffa102916a>] lod_declare_xattr_set_lmv+0xe8a/0x28d0 [lod]
<4>[ 4377.643702]  [<ffffffffa0b277f8>] ? qsd_op_begin+0xb8/0xb70 [lquota]
<4>[ 4377.643855]  [<ffffffffa102ade7>] lod_dir_striping_create_internal+0x237/0x1b40 [lod]
<4>[ 4377.644119]  [<ffffffffa123d331>] ? osd_declare_quota+0x1c1/0x2d0 [osd_zfs]
<4>[ 4377.644293]  [<ffffffffa124314a>] ? osd_declare_object_create+0x33a/0x440 [osd_zfs]
<4>[ 4377.644543]  [<ffffffffa123edb1>] ? osd_object_write_unlock+0x61/0x70 [osd_zfs]
<4>[ 4377.644791]  [<ffffffffa102c951>] lod_declare_object_create+0x261/0x3e0 [lod]
<4>[ 4377.644954]  [<ffffffffa0f1a846>] mdd_declare_object_create_internal+0x116/0x340 [mdd]
<4>[ 4377.645211]  [<ffffffffa0f1605e>] mdd_create+0x68e/0x1730 [mdd]
<4>[ 4377.645366]  [<ffffffffa0f779cc>] ? mdt_version_save+0x8c/0x1a0 [mdt]
<4>[ 4377.645521]  [<ffffffffa0f7bdc1>] mdt_reint_create+0xbf1/0xd40 [mdt]
<4>[ 4377.645689]  [<ffffffffa0795b10>] ? lu_ucred+0x20/0x30 [obdclass]
<4>[ 4377.645854]  [<ffffffffa0f57d15>] ? mdt_ucred+0x15/0x20 [mdt]
<4>[ 4377.646102]  [<ffffffffa0f7282c>] ? mdt_root_squash+0x2c/0x3f0 [mdt]
<4>[ 4377.646371]  [<ffffffffa09b98b2>] ? __req_capsule_get+0x162/0x6d0 [ptlrpc]
<4>[ 4377.646533]  [<ffffffffa0f76add>] mdt_reint_rec+0x5d/0x200 [mdt]
<4>[ 4377.646692]  [<ffffffffa0f5c64b>] mdt_reint_internal+0x4cb/0x7a0 [mdt]
<4>[ 4377.646847]  [<ffffffffa0f5cdbb>] mdt_reint+0x6b/0x120 [mdt]
<4>[ 4377.647009]  [<ffffffffa09f312e>] tgt_request_handle+0x8be/0x1000 [ptlrpc]
<4>[ 4377.647185]  [<ffffffffa09a4a34>] ptlrpc_main+0xdf4/0x1940 [ptlrpc]
<4>[ 4377.647349]  [<ffffffffa09a3c40>] ? ptlrpc_main+0x0/0x1940 [ptlrpc]
<4>[ 4377.647500]  [<ffffffff8109ce4e>] kthread+0x9e/0xc0
<4>[ 4377.647636]  [<ffffffff8100c24a>] child_rip+0xa/0x20
<4>[ 4377.647773]  [<ffffffff8109cdb0>] ? kthread+0x0/0xc0
<4>[ 4377.647909]  [<ffffffff8100c240>] ? child_rip+0x0/0x20
<4>[ 4377.648053] 
<0>[ 4377.649299] Kernel panic - not syncing: LBUG

I guess this is a bit similar to LU-2911 and LU-2958,except I do not have any olf FS configs or anthing of the nature, everything is freshly reformatted before the test starts.

Issue Links

• duplicates: LU-6154 striped directory on ZFS

Comments

[Oleg Drokin]

I just tested and ldiskfs does not crash in this case, so it's a zfs-only issue apparently.

[Di Wang]

I believe this can be fixed by the patch in LU-6154 http://review.whamcloud.com/#/c/13518/

[Di Wang]

duplicate with LU-6154