[LU-1708] ASSERTION(__v > 0 && __v < ((int)0x5a5a5a5a5a5a5a5a))

• Status: Resolved
• Project: Lustre
• Affects Version/s: Lustre 2.1.2
• Type: Bug
• Priority: Minor
• Reporter: Mahmoud Hanafi
• Assignee: Bob Glossman (Inactive)
• Resolution: Duplicate
• Severity: 3

Description

Hit assertion after IB fabric work and SM restart.
Lustre: Skipped 59 previous similar messages^M
LustreError: 6831:0:(mgs_handler.c:782:mgs_handle()) MGS handle cmd=250 rc=-114^M
LustreError: 6832:0:(mgs_handler.c:782:mgs_handle()) MGS handle cmd=250 rc=-114^M
LustreError: 6831:0:(mgs_handler.c:782:mgs_handle()) Skipped 46 previous similar messages^M
LustreError: 6832:0:(mgs_handler.c:782:mgs_handle()) Skipped 46 previous similar messages^M
Lustre: 6369:0:(ldlm_lib.c:947:target_handle_connect()) nbp1-MDT0000: connection from 24c3494a-96ad-f0c6-3887-efa3caaaf377@10.151.32.247@o2ib t377954918473 exp (null) cur 1344015983 last 0^M
Lustre: nbp1-MDT0000: denying duplicate export for 24c3494a-96ad-f0c6-3887-efa3caaaf377, -114^M
Lustre: Skipped 44 previous similar messages^M
Lustre: 6369:0:(ldlm_lib.c:947:target_handle_connect()) Skipped 184 previous similar messages^M
LustreError: 5862:0:(obd_class.h:501:obd_set_info_async()) obd_set_info_async: dev 0 no operation^M
LustreError: 5862:0:(obd_class.h:501:obd_set_info_async()) Skipped 36 previous similar messages^M
Lustre: MGS: Client 0189e00e-6900-6c13-4b08-b178fca62eec (at 10.151.52.204@o2ib) reconnecting^M
Lustre: Skipped 232 previous similar messages^M
LustreError: 6369:0:(mdt_handler.c:2792:mdt_recovery()) operation 400 on unconnected MDS from 12345-10.151.36.209@o2ib^M
LustreError: 6369:0:(mdt_handler.c:2792:mdt_recovery()) Skipped 698 previous similar messages^M
Lustre: nbp1-MDT0000: Export ffff88080a5c5c00 already connecting from 10.151.36.209@o2ib^M
Lustre: Skipped 112 previous similar messages^M
Lustre: 5862:0:(ldlm_lib.c:947:target_handle_connect()) MGS: connection from 84bf4c4b-a930-7713-e68b-5d60d626e2d0@10.151.28.217@o2ib t0 exp (null) cur 1344015999 last 0^M
Lustre: nbp1-MDT0000: denying duplicate export for c40bbbf1-1075-f61f-614a-3eb214d31ac0, -114^M
Lustre: Skipped 172 previous similar messages^M
Lustre: 5862:0:(ldlm_lib.c:947:target_handle_connect()) Skipped 637 previous similar messages^M
LustreError: 6820:0:(obd_class.h:501:obd_set_info_async()) obd_set_info_async: dev 0 no operation^M
LustreError: 6820:0:(obd_class.h:501:obd_set_info_async()) Skipped 122 previous similar messages^M
LustreError: 6809:0:(genops.c:970:class_import_put()) ASSERTION(__v > 0 && __v < ((int)0x5a5a5a5a5a5a5a5a)) failed: value: 0^M
LustreError: 6809:0:(genops.c:970:class_import_put()) LBUG^M
Pid: 6809, comm: ll_mgs_06^M
^M
Call Trace:^M
[<ffffffffa0605855>] libcfs_debug_dumpstack+0x55/0x80 [libcfs]^M
[<ffffffffa0605e95>] lbug_with_loc+0x75/0xe0 [libcfs]^M
[<ffffffffa06d1678>] class_import_put+0x298/0x330 [obdclass]^M
^M
Entering kdb (current=0xffff8811f6c32a80, pid 6809) on processor 13 Oops: (null)^M

DUPLICATE OF ORI0710

Comments

[Peter Jones]

Bob

Could you please look into this one?

Thanks

Peter

[Bob Glossman (Inactive)]

There is a fix in progress that may help, http://review.whamcloud.com/#change,3538. This hasn't been fully tested and landed yet so if your problem isn't urgent you may want to wait.

[Bob Glossman (Inactive)]

This problem also has features similar to LU-1432, so the fix from http://review.whamcloud.com/#change,3244 might also be helpful.
That fix has already landed in b2_1 for the next 2.1.x release, so it's probably safe to use.

[Jay Lan (Inactive)]

This is actually a dup of ORI-710. ORI-710 was closed on Oct 7 with patch landed to master branch under LU-2007. I cherry-picked the LU-2007 patch (bf625a1) to our 2.1.3 branch and it was a clean cherry-pick.

This LU can be closed when LU-2007 patch lands on b2_1 branch.

[Mahmoud Hanafi]

This can be closed

[Peter Jones]

LU-2007 was included in 2.1.4 onwards