[LU-7223] Failover - mmp: MMP was not enabled on /dev/lvm-Role_OSS/P1 on ost1

• Status: Resolved
• Project: Lustre
• Affects Version/s: Lustre 2.8.0
• Fix Version/s: Lustre 2.8.0
• Type: Bug
• Priority: Minor
• Reporter: Maloo
• Assignee: Niu Yawei (Inactive)
• Resolution: Fixed

Environment

Client: 2.7.60, SLES 11.3
Server: 2.7.60 , RHEL 6.7

• Severity: 3

Description

This issue was created by maloo for Saurabh Tandan <[e-mail]>

No subtests ran. suite-log shows following issue:

modules unloaded.
excepting tests: 
CMD: onyx-42vm3 dumpe2fs -h /dev/lvm-Role_MDS/P1 | grep mmp
onyx-42vm3: dumpe2fs 1.42.13.wc3 (28-Aug-2015)
Filesystem features:      has_journal ext_attr resize_inode dir_index filetype mmp flex_bg dirdata sparse_super large_file huge_file uninit_bg dir_nlink quota
CMD: onyx-42vm4 dumpe2fs -h /dev/lvm-Role_OSS/P1 | grep mmp
onyx-42vm4: dumpe2fs 1.42.13.wc3 (28-Aug-2015)
 mmp : @@@@@@ FAIL: MMP was not enabled on /dev/lvm-Role_OSS/P1 on ost1 

This issue relates to the following test suite run: https://testing.hpdd.intel.com/test_sets/bc9071ca-632c-11e5-afa7-5254006e85c2.

Comments

[Peter Jones]

Niu

Could you please look into this issue?

Thanks

Peter

[Saurabh Tandan (Inactive)]

master, build# 3264, 2.7.64 tag
Hard Failover: EL6.7 Server/Client
https://testing.hpdd.intel.com/test_sets/80953fd8-9edd-11e5-87a9-5254006e85c2

[Saurabh Tandan (Inactive)]

This test has failed with the same issue around 25 times in the span of past 30 days.

[Saurabh Tandan (Inactive)]

master, build# 3266, 2.7.64 tag
Hard Failover: EL6.7 Server/SLES11 SP3 Clients
https://testing.hpdd.intel.com/test_sets/bca45d52-a080-11e5-85ed-5254006e85c2

Hard Failover: EL7 Server/SLES11 SP3 Client
https://testing.hpdd.intel.com/test_sets/a8a78e54-a077-11e5-8d69-5254006e85c2

[Sarah Liu]

The failure blocks the MMP suites (10 subtests) of being run.

[Niu Yawei (Inactive)]

Can't tell why the 'mmp' not enabled on OST device, I pushed a debug patch to collect more information when the test failed: http://review.whamcloud.com/17628

[Joseph Gmitter (Inactive)]

Hi Niu,
Is there any further info available from the debug patch?
Thanks.
Joe

[Niu Yawei (Inactive)]

No, the debug patch passed test on maloo, I think we'd commit this debug patch so that it can be triggered more likely. I'm going to update the patch and try to get it reviewed and landed.

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch http://review.whamcloud.com/17628/
Subject: LU-7223 tests: print more information when mmp.sh failed
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: eb6cd4804d65dda1b6ea4a1289cc01647d03a47a

[Niu Yawei (Inactive)]

The landed patch is a debug patch.

[Andreas Dilger]

Looking at the most recent test failure https://testing.hpdd.intel.com/test_sets/56d8d2dc-b502-11e5-8114-5254006e85c2 it seems that MMP is enabled but the check is missing this feature?

[Niu Yawei (Inactive)]

Andreas, the displayed features including mmp is from MDS, but not OSS. That test seems doesn't have the debug patch included.

I found one failure which has debug patch included, https://testing.hpdd.intel.com/test_logs/d3a29f9e-b4be-11e5-8114-5254006e85c2/show_text , I think the script checking the mmp feature by looking at the failover_HOST variable isn't reliable (seems the variable could be exported even if no failover being used), I'm going to change it to looking at the features on device directly.

[Gerrit Updater]

Niu Yawei ([e-mail]) uploaded a new patch: http://review.whamcloud.com/17884
Subject: LU-7223 test: check mmp feature properly in mmp.sh
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 4400c997b37e4c365c9d35ebe6b710509f0323f0

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch http://review.whamcloud.com/17884/
Subject: LU-7223 test: check mmp feature properly in mmp.sh
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 4d941c27bd18528a150ae5a262388e2a089f2165

[Peter Jones]

Landed for 2.8