[LU-2725] ptlrpc_server_hpreq_init()) A SSERTION( rc == 0 || rc == 1 ) failed

• Type: Bug
• Resolution: Fixed
• Priority: Blocker
• Fix Version/s: Lustre 2.4.0
• Affects Version/s: Lustre 2.4.0
• Severity: 3
• Assignee: WC Triage
• Reporter: Oleg Drokin

Description

Hit this in racer, but I suspect it will hit in other workloads when hpreqs are happening.

[ 3577.720282] LustreError: 12694:0:(service.c:1512:ptlrpc_server_hpreq_init()) A
SSERTION( rc == 0 || rc == 1 ) failed: 
[ 3577.720820] LustreError: 12694:0:(service.c:1512:ptlrpc_server_hpreq_init()) L
BUG
[ 3577.721172] Pid: 12694, comm: ll_ost_io00_007
[ 3577.721320] 
[ 3577.721320] Call Trace:
[ 3577.721675]  [<ffffffffa0449915>] libcfs_debug_dumpstack+0x55/0x80 [libcfs]
[ 3577.721976]  [<ffffffffa0449f17>] lbug_with_loc+0x47/0xb0 [libcfs]
[ 3577.722287]  [<ffffffffa079dc9a>] ptlrpc_main+0x15ca/0x17f0 [ptlrpc]
[ 3577.722591]  [<ffffffffa079c6d0>] ? ptlrpc_main+0x0/0x17f0 [ptlrpc]
[ 3577.722882]  [<ffffffff8100c14a>] child_rip+0xa/0x20
[ 3577.723169]  [<ffffffffa079c6d0>] ? ptlrpc_main+0x0/0x17f0 [ptlrpc]
[ 3577.723473]  [<ffffffffa079c6d0>] ? ptlrpc_main+0x0/0x17f0 [ptlrpc]
[ 3577.723758]  [<ffffffff8100c140>] ? child_rip+0x0/0x20
[ 3577.724014] 
[ 3577.762250] Kernel panic - not syncing: LBUG

The rc value is 2.
The culprit seems to be ost_rw_hpreq_check returning number of locks matched instead of 1 to show the matching happened and NRS changes actually demand either 0 or 1 return value

Comments

[Oleg Drokin]

patch landed

[Oleg Drokin]

Patch in http://review.whamcloud.com/5233