[LU-12963] recovery-small test 110k hangs

Related — Tue, 12 Nov 2019 21:33:43 +0000

recovery-small test_110k hangs. It looks like these hangs started on 21 JULY 2019 for Lustre 2.12.56.17 and only seen on master/b2_13.

Looking at the console logs for the hang at https://testing.whamcloud.com/test_sets/b732f8c4-e37a-11e9-add9-52540065bddc, we see the following:

client1 (vm11)

[18823.080898] Lustre: DEBUG MARKER: == recovery-small test 110k: FID_QUERY failed during recovery ======================================== 15:24:43 (1569770683)
[18826.019012] LustreError: 11-0: lustre-MDT0001-mdc-ffff986e5f303800: operation mds_statfs to node 10.9.4.165@tcp failed: rc = -107
[18826.020320] LustreError: Skipped 4 previous similar messages
[22427.043134] sysrq: SysRq : Changing Loglevel
[22427.043724] sysrq: Loglevel set to 8
[22427.046389] sysrq: SysRq : Show State
[22427.046821]   task                        PC stack   pid father
[22427.047454] systemd         S    0     1      0 0x00000000
[22427.048036] Call Trace:
[22427.048363]  ? __schedule+0x254/0x840
[22427.048753]  schedule+0x28/0x70
[22427.049116]  schedule_hrtimeout_range_clock+0x18d/0x1a0
[22427.049667]  ? ep_scan_ready_list.constprop.22+0x20c/0x230
[22427.050253]  ep_poll+0x3c8/0x410
[22427.050622]  ? wake_up_q+0x70/0x70
[22427.050994]  do_epoll_wait+0xb0/0xd0
[22427.051394]  __x64_sys_epoll_wait+0x1a/0x20
[22427.051842]  do_syscall_64+0x5b/0x1b0
[22427.052260]  entry_SYSCALL_64_after_hwframe+0x65/0xca
[22427.052789] RIP: 0033:0x7f49dea94d97
[22427.053197] Code: Bad RIP value.
[22427.053545] RSP: 002b:00007fffe0967b50 EFLAGS: 00000293 ORIG_RAX: 00000000000000e8
[22427.054327] RAX: ffffffffffffffda RBX: 0000000000000004 RCX: 00007f49dea94d97
[22427.055061] RDX: 000000000000005c RSI: 00007fffe0967b90 RDI: 0000000000000004
[22427.055777] RBP: 00007fffe0967b90 R08: 0000000000000000 R09: 7465677261742e79
[22427.056507] R10: 00000000ffffffff R11: 0000000000000293 R12: 000000000000005c
[22427.057238] R13: 00000000ffffffff R14: 00007fffe0967b90 R15: 0000000000000001
[22427.057954] kthreadd        S    0     2      0 0x80000000
[22427.058521] Call Trace:
[22427.058804]  ? __schedule+0x254/0x840
[22427.059207]  schedule+0x28/0x70
[22427.059550]  kthreadd+0x2e4/0x300
[22427.059912]  ? kthread_create_on_cpu+0xa0/0xa0
[22427.060397]  ret_from_fork+0x35/0x40
[22427.060794] rcu_gp          I    0     3      2 0x80000000
[22427.061372] Workqueue:            (null) (rcu_gp)
[22427.061855] Call Trace:
[22427.062147]  ? __schedule+0x254/0x840
[22427.062544]  schedule+0x28/0x70
[22427.062886]  rescuer_thread+0x2ee/0x340
[22427.063319]  ? flush_rcu_work+0x30/0x30
[22427.063721]  kthread+0x112/0x130
[22427.064084]  ? kthread_bind+0x30/0x30
[22427.064476]  ret_from_fork+0x35/0x40
[22427.064854] rcu_par_gp      I    0     4      2 0x80000000
[22427.065429] Call Trace:
[22427.065708]  ? __schedule+0x254/0x840
[22427.066113]  schedule+0x28/0x70
[22427.066455]  rescuer_thread+0x2ee/0x340
[22427.066855]  ? flush_rcu_work+0x30/0x30
[22427.067282]  kthread+0x112/0x130
[22427.067637]  ? kthread_bind+0x30/0x30
[22427.068038]  ret_from_fork+0x35/0x40
…
[22428.500219] kworker/1:1     D    0 16467      2 0x80000080
[22428.500791] Workqueue: events key_garbage_collector
[22428.501318] Call Trace:
[22428.501596]  ? __schedule+0x254/0x840
[22428.501993]  schedule+0x28/0x70
[22428.502343]  schedule_timeout+0x26d/0x390
[22428.502766]  ? check_preempt_curr+0x62/0x90
[22428.503222]  wait_for_completion+0x11f/0x190
[22428.503667]  ? wake_up_q+0x70/0x70
[22428.504050]  __wait_rcu_gp+0x12c/0x160
[22428.504453]  synchronize_rcu+0x76/0x80
[22428.504853]  ? __call_rcu+0x280/0x280
[22428.505260]  ? __bpf_trace_rcu_utilization+0x10/0x10
[22428.505773]  key_garbage_collector+0x1d7/0x3e0
[22428.506268]  process_one_work+0x1a7/0x360
[22428.506690]  worker_thread+0x30/0x390
[22428.507102]  ? pwq_unbound_release_workfn+0xd0/0xd0
[22428.507608]  kthread+0x112/0x130
[22428.507958]  ? kthread_bind+0x30/0x30
[22428.508363]  ret_from_fork+0x35/0x40

client2 (vm12)

[18827.149580] Lustre: DEBUG MARKER: == recovery-small test 110k: FID_QUERY failed during recovery ======================================== 15:24:43 (1569770683)
[22433.833049] sysrq: SysRq : Changing Loglevel
[22433.833608] sysrq: Loglevel set to 8
[22433.838786] sysrq: SysRq : Show State
[22433.839253]   task                        PC stack   pid father
[22433.839871] systemd         S    0     1      0 0x00000000
[22433.840457] Call Trace:
[22433.840780]  ? __schedule+0x254/0x840
[22433.841198]  schedule+0x28/0x70
[22433.841551]  schedule_hrtimeout_range_clock+0x18d/0x1a0
[22433.842134]  ? ep_scan_ready_list.constprop.22+0x20c/0x230
[22433.842706]  ep_poll+0x3c8/0x410
[22433.843101]  ? wake_up_q+0x70/0x70
[22433.843477]  do_epoll_wait+0xb0/0xd0
[22433.843864]  __x64_sys_epoll_wait+0x1a/0x20
[22433.844330]  do_syscall_64+0x5b/0x1b0
[22433.844735]  entry_SYSCALL_64_after_hwframe+0x65/0xca
[22433.845288] RIP: 0033:0x7fd7332c2d97
[22433.845678] Code: Bad RIP value.
[22433.846045] RSP: 002b:00007ffdb0a7b080 EFLAGS: 00000293 ORIG_RAX: 00000000000000e8
[22433.846810] RAX: ffffffffffffffda RBX: 0000000000000004 RCX: 00007fd7332c2d97
[22433.847548] RDX: 0000000000000059 RSI: 00007ffdb0a7b0c0 RDI: 0000000000000004
[22433.848290] RBP: 00007ffdb0a7b0c0 R08: 0000000000000000 R09: 7465677261742e79
[22433.849032] R10: 00000000ffffffff R11: 0000000000000293 R12: 0000000000000059
[22433.849755] R13: 00000000ffffffff R14: 00007ffdb0a7b0c0 R15: 0000000000000001
[22433.850497] kthreadd        S    0     2      0 0x80000000
[22433.851081] Call Trace:
[22433.851364]  ? __schedule+0x254/0x840
[22433.851756]  schedule+0x28/0x70
[22433.852118]  kthreadd+0x2e4/0x300
[22433.852494]  ? kthread_create_on_cpu+0xa0/0xa0
[22433.852965]  ret_from_fork+0x35/0x40
[22433.853384] rcu_gp          I    0     3      2 0x80000000
[22433.853959] Call Trace:
[22433.854259]  ? __schedule+0x254/0x840
[22433.854655]  schedule+0x28/0x70
[22433.855020]  rescuer_thread+0x2ee/0x340
[22433.855443]  ? flush_rcu_work+0x30/0x30
[22433.855856]  kthread+0x112/0x130
[22433.856239]  ? kthread_bind+0x30/0x30
[22433.856635]  ret_from_fork+0x35/0x40
…
[ 8826.737348] kworker/1:1     D    0  5439      2 0x80000080
[ 8826.737911] Workqueue: events key_garbage_collector
[ 8826.738582] Call Trace:
[ 8826.738989]  ? __schedule+0x254/0x840
[ 8826.739599]  ? pcpu_next_unpop+0x37/0x50
[ 8826.740216]  schedule+0x28/0x70
[ 8826.740713]  schedule_timeout+0x26d/0x390
[ 8826.741367]  ? _cond_resched+0x15/0x30
[ 8826.741859]  ? cpumask_next+0x17/0x20
[ 8826.742372]  ? kmem_cache_open+0x3f0/0x430
[ 8826.742804]  wait_for_completion+0x11f/0x190
[ 8826.743451]  ? wake_up_q+0x70/0x70
[ 8826.743999]  __wait_rcu_gp+0x12c/0x160
[ 8826.744654]  synchronize_rcu+0x76/0x80
[ 8826.745261]  ? __call_rcu+0x280/0x280
[ 8826.745857]  ? __bpf_trace_rcu_utilization+0x10/0x10
[ 8826.746413]  key_garbage_collector+0x1d7/0x3e0
[ 8826.747031]  process_one_work+0x1a7/0x360
[ 8826.747687]  worker_thread+0x1fa/0x390
[ 8826.748294]  ? pwq_unbound_release_workfn+0xd0/0xd0
[ 8826.748839]  kthread+0x112/0x130
[ 8826.749434]  ? kthread_bind+0x30/0x30
[ 8826.750218]  ret_from_fork+0x35/0x40

MDS1,3 (vm6)

[18819.270132] Lustre: DEBUG MARKER: == recovery-small test 110k: FID_QUERY failed during recovery ======================================== 15:24:43 (1569770683)
[18822.297574] LustreError: 11-0: lustre-MDT0001-osp-MDT0002: operation mds_statfs to node 10.9.4.165@tcp failed: rc = -107
[18822.299461] LustreError: Skipped 6 previous similar messages

MDS2,4 (vm7)

[18824.196956] Lustre: DEBUG MARKER: == recovery-small test 110k: FID_QUERY failed during recovery ======================================== 15:24:43 (1569770683)
[18824.510121] Lustre: DEBUG MARKER: grep -c /mnt/lustre-mds2' ' /proc/mounts || true
[18825.124512] Lustre: DEBUG MARKER: umount -d /mnt/lustre-mds2
[18825.423041] Lustre: Failing over lustre-MDT0001
[18825.423922] Lustre: Skipped 7 previous similar messages
[18825.657887] Lustre: server umount lustre-MDT0001 complete
[18825.658935] Lustre: Skipped 7 previous similar messages
[18826.212991] Lustre: DEBUG MARKER: lsmod | grep lnet > /dev/null &&
[18826.212991] lctl dl | grep ' ST ' || true
[18826.797370] Lustre: DEBUG MARKER: modprobe dm-flakey;
[18826.797370] 			 dmsetup targets | grep -q flakey
[18826.868811] LustreError: 137-5: lustre-MDT0001_UUID: not available for connect from 10.9.4.165@tcp (no target). If you are running an HA pair check that the target is mounted on the other server.
[18826.871737] LustreError: Skipped 187 previous similar messages
[18903.428904] LustreError: 137-5: lustre-MDT0001_UUID: not available for connect from 10.9.4.165@tcp (no target). If you are running an HA pair check that the target is mounted on the other server.
[18903.432265] LustreError: Skipped 197 previous similar messages
[19053.669128] LustreError: 137-5: lustre-MDT0001_UUID: not available for connect from 10.9.4.165@tcp (no target). If you are running an HA pair check that the target is mounted on the other server.
[19053.672181] LustreError: Skipped 387 previous similar messages
[19354.148751] LustreError: 137-5: lustre-MDT0001_UUID: not available for connect from 10.9.4.165@tcp (no target). If you are running an HA pair check that the target is mounted on the other server.
[19354.152173] LustreError: Skipped 777 previous similar messages
[19954.814825] LustreError: 137-5: lustre-MDT0001_UUID: not available for connect from 10.9.4.118@tcp (no target). If you are running an HA pair check that the target is mounted on the other server.
[19954.817893] LustreError: Skipped 1553 previous similar messages
[20556.069346] LustreError: 137-5: lustre-MDT0001_UUID: not available for connect from 10.9.4.165@tcp (no target). If you are running an HA pair check that the target is mounted on the other server.
[20556.072539] LustreError: Skipped 1555 previous similar messages
[21157.029130] LustreError: 137-5: lustre-MDT0001_UUID: not available for connect from 10.9.4.165@tcp (no target). If you are running an HA pair check that the target is mounted on the other server.
[21157.032971] LustreError: Skipped 1552 previous similar messages
[21757.094270] LustreError: 137-5: lustre-MDT0001_UUID: not available for connect from 10.9.4.118@tcp (no target). If you are running an HA pair check that the target is mounted on the other server.
[21757.097259] LustreError: Skipped 1553 previous similar messages
[22358.949018] LustreError: 137-5: lustre-MDT0001_UUID: not available for connect from 10.9.4.165@tcp (no target). If you are running an HA pair check that the target is mounted on the other server.
[22358.952110] LustreError: Skipped 1555 previous similar messages
[22443.966949] SysRq : Changing Loglevel
[22443.967688] Loglevel set to 8

We have several instances of this hang:
https://testing.whamcloud.com/test_sets/74cea80c-03e1-11ea-b934-52540065bddc
https://testing.whamcloud.com/test_sets/304f5830-03f9-11ea-a9d7-52540065bddc

Whamcloud Community JIRA

[LU-12963] recovery-small test 110k hangs