[LU-2070] lu_object.c:1114:lu_device_fini()) ASSERTIO N( cfs_atomic_read(&d->ld_ref) == 0 ) failed: Refcount is 6

• Type: Bug
• Resolution: Duplicate
• Priority: Blocker
• Affects Version/s: Lustre 2.4.0
• Environment: single node vm, 10G RAM, 8 cores
• Severity: 3
• Assignee: WC Triage
• Reporter: Oleg Drokin

Description

This is the second time I hit this problem in sanity 132:

== sanity test 132: som avoids glimpse rpc == 21:26:31 (1349141191)
====> SOM is disabled, 1 glimpse RPC occured
remounting...
Stopping clients: rhel6.localnet /mnt/lustre (opts:)
Stopping client rhel6.localnet /mnt/lustre opts:
Stopping clients: rhel6.localnet /mnt/lustre2 (opts:)
Stopping /mnt/mds1 (opts:-f) on rhel6.localnet

Message from syslogd@rhel6 at Oct  1 21:26:43 ...
 kernel:[ 3098.596851] LustreError: 21090:0:(lu_object.c:1114:lu_device_fini()) ASSERTION( cfs_atomic_read(&d->ld_ref) == 0 ) failed: Refcount is 6

the trace is

[ 3091.895067] Lustre: Unmounted lustre-client
[ 3098.583444] LustreError: 8541:0:(osp_dev.c:823:osp_import_event()) lustre-OST0000-osc-MDT0000: unsupported import event: 0x808003
[ 3098.585314] LustreError: 8541:0:(osp_dev.c:823:osp_import_event()) Skipped 2 previous similar messages
[ 3098.596851] LustreError: 21090:0:(lu_object.c:1114:lu_device_fini()) ASSERTIO
N( cfs_atomic_read(&d->ld_ref) == 0 ) failed: Refcount is 6[ 3098.598740] LustreError: 21090:0:(lu_object.c:1114:lu_device_fini()) LBUG
[ 3098.599767] Pid: 21090, comm: obd_zombid
[ 3098.600380][ 3098.600381] Call Trace:
[ 3098.601082]  [<ffffffffa0ce8915>] libcfs_debug_dumpstack+0x55/0x80 [libcfs][ 3098.603116]  [<ffffffffa0ce8f27>] lbug_with_loc+0x47/0xb0 [libcfs]
[ 3098.604397]  [<ffffffffa031ec6c>] lu_device_fini+0xcc/0xd0 [obdclass]
[ 3098.605835]  [<ffffffffa032535e>] dt_device_fini+0xe/0x10 [obdclass]
[ 3098.606843]  [<ffffffffa0241736>] osp_device_free+0xd6/0x1f0 [osp]
[ 3098.607807]  [<ffffffffa0303a4d>] class_decref+0x46d/0x590 [obdclass]
[ 3098.608938]  [<ffffffffa02e0784>] obd_zombie_impexp_cull+0x314/0x620 [obdclass]
[ 3098.610081]  [<ffffffffa02e0b55>] obd_zombie_impexp_thread+0xc5/0x1c0 [obdclass]
[ 3098.611235]  [<ffffffff81057d60>] ? default_wake_function+0x0/0x20
[ 3098.612195]  [<ffffffffa02e0a90>] ? obd_zombie_impexp_thread+0x0/0x1c0 [obdclass]
[ 3098.613570]  [<ffffffff8100c14a>] child_rip+0xa/0x20
[ 3098.614363]  [<ffffffffa02e0a90>] ? obd_zombie_impexp_thread+0x0/0x1c0 [obdclass]
[ 3098.615532]  [<ffffffffa02e0a90>] ? obd_zombie_impexp_thread+0x0/0x1c0 [obdclass]
[ 3098.616696]  [<ffffffff8100c140>] ? child_rip+0x0/0x20
[ 3098.617687]
[ 3098.618580] Kernel panic - not syncing: LBUG

Issue Links

• is related to: LU-1480 failure on replay-single test_74: ASSERTION( cfs_atomic_read(&d->ld_ref) == 0 ) failed: Refcount is 1