[LU-12616] MDS node crashed LustreError: 23042:0:(mdt_handler.c:5135:mdt_init0()) ASSERTION( info != ((void *)0) )

• Type: Bug
• Resolution: Fixed
• Priority: Major
• Fix Version/s: Lustre 2.13.0
• Labels: LTS12 patch
• Severity: 3
• Assignee: Alexander Boyko
• Reporter: Alexander Boyko

Description

[ 1832.493775] LNet: 26562:0:(socklnd_cb.c:425:ksocknal_txlist_done()) Deleting packet type 1 len 520 172.18.1.3@tcp->172.18.1.4@tcp
[ 1870.590047] LustreError: 20610:0:(mgc_request.c:249:do_config_log_add()) MGC172.18.1.3@tcp: failed processing log, type 4: rc = -110
[ 1882.605517] LustreError: 23042:0:(mdt_handler.c:5135:mdt_init0()) ASSERTION( info != ((void *)0) ) failed:
[ 1882.616438] LustreError: 23042:0:(mdt_handler.c:5135:mdt_init0()) LBUG  

 The dk log shows the next steps

started cleanup of MDT01
started cleanup of MDT00
finished cleanup of MDT01, and cleanup of MDS also
started MDT01 mount + setup of MDS
finished setup of MDS
finished cleanup of MDT00, and cleanup of MDS also
asserted during MDT01 initialization
The main problem is MDS was stopped during MDT01 mount. It looks like wrong cleanup of MDS.

Comments

[Gerrit Updater]

Minh Diep ([e-mail]) uploaded a new patch: https://review.whamcloud.com/39157
Subject: LU-12616 obclass: fix MDS start/stop race
Project: fs/lustre-release
Branch: b2_12
Current Patch Set: 1
Commit: 8c757edba00b5fd6ddf76eb41b41fd398e95eb66

[Peter Jones]

Landed for 2.13

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch https://review.whamcloud.com/35652/
Subject: LU-12616 obclass: fix MDS start/stop race
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 3cce65712d94cffe8f1626545845b95b88aef672

[James A Simmons]

I think I might of caused this. Can you try patch https://review.whamcloud.com/#/c/34718/

[Gerrit Updater]

Alexandr Boyko ([e-mail]) uploaded a new patch: https://review.whamcloud.com/35652
Subject: LU-12616 obclass: fix MDS start/stop race
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 347718de9bfcd759949a0a56221ae5b75afd02dd