[LU-901] LustreError: 25196:0:(client.c:1696:ptlrpc_check_set()) LBUG

• Status: Resolved
• Project: Lustre
• Type: Bug
• Priority: Major
• Reporter: Alexey Lyashkov
• Assignee: Emoly Liu
• Resolution: Cannot Reproduce
• Labels: patch
• Severity: 3

Description

when cmd3 branch updated from 1.x branch, it lost a one fix, which hit recently.
https://bugzilla.lustre.org/show_bug.cgi?id=11544

issue hit when client tried to resend a bulk requests after recovery.
...
00000100:00000001:0.0:1323101799.643655:0:23526:0:(events.c:184:client_bulk_callback()) Process entered
00000100:00020000:0.0:1323101799.643655:0:23526:0:(events.c:195:client_bulk_callback()) event type 0, status -5, desc ffff88004dedaa00
00000100:00000001:0.0:1323101799.645477:0:23526:0:(events.c:223:client_bulk_callback()) Process leaving
...
00000100:00000001:0.0:1323101799.651005:0:23528:0:(client.c:1416:ptlrpc_check_set()) Process entered
00000100:00020000:0.0:1323101799.651007:0:23528:0:(client.c:1695:ptlrpc_check_set()) @@@ bulk transfer failed req@ffff880053974400 x1387372579127387/t1983(1983) o4->testfs-OST0007-osc-ffff88005341d800@192.168.2.13@tcp:6/4 lens 456/416 e 0 to 1 dl 1323101749 ref 2 fl Bulk:RXS/4/0 rc -11/0
00000000:00040000:0.0:1323101799.655434:0:23528:0:(client.c:1696:ptlrpc_check_set()) LBUG
...

I can attach a full debug log, if need.

Issue Links

• is duplicated by: LU-2260 2.1.3 Client LBUG

Comments

[Andreas Dilger]

Shadow, can you please submit the patch from bugzilla to Gerrit so that it can begin testing.

[Alexey Lyashkov]

i submit minimal patch into LU-1039, but reserve that ticket to more better fix.
but it's fix need to make big cleanup in check_set() function to make FSM more clean and state changes need clarified correctly.

[Andreas Dilger]

Shadow, are you planning to do anything with this bug? Otherwise, please close it.

[Alexey Lyashkov]

yes, i have plan. patch partially finished - and need resolve one issue with imp in flight requests.
I think it's will be ready before LUG.

[Alexey Lyashkov]

remote: New Changes:
remote: http://review.whamcloud.com/4092

[Nathan Rutman]

Xyratex-bug-id: MRP-355

[Andreas Dilger]

Close old ticket.