[LU-6637] sanity-scrub test_9: Got speed 158, expected more than 168

• Status: Resolved
• Project: Lustre
• Type: Bug
• Priority: Minor
• Reporter: Maloo
• Assignee: WC Triage
• Resolution: Cannot Reproduce
• Labels: dne2
• Severity: 3

Description

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

This issue relates to the following test suite run: https://testing.hpdd.intel.com/test_sets/1781a240-022d-11e5-9636-5254006e85c2.

The sub-test test_9 failed with the following error:

(11) Got speed 158, expected more than 168

Although I only saw this failure on my DNE2 test, but my patch did not touch these area at all. So it is probably related with sth else.

sanity-scrub test_9: @@@@@@ FAIL: (11) Got speed 158, expected more than 168 
  Trace dump:
  = /usr/lib64/lustre/tests/test-framework.sh:4727:error_noexit()
  = /usr/lib64/lustre/tests/test-framework.sh:4758:error()
  = /usr/lib64/lustre/tests/sanity-scrub.sh:848:test_9()
  = /usr/lib64/lustre/tests/test-framework.sh:5020:run_one()
  = /usr/lib64/lustre/tests/test-framework.sh:5057:run_one_logged()
  = /usr/lib64/lustre/tests/test-framework.sh:4907:run_test()
  = /usr/lib64/lustre/tests/sanity-scrub.sh:860:main()
Dumping lctl log to /logdir/test_logs/2015-05-24/lustre-reviews-el6_6-x86_64--review-dne-part-2--2_9_1__32460__-70316177025900-064318/sanity-scrub.test_9.*.1432462607.log

Issue Links

• is related to: LU-6831 The ticket for tracking all DNE2 bugs

Comments

[Di Wang]

Fan Yong: could you please advise here? Thanks!

[nasf (Inactive)]

The test failure means the scrub speed in not as good as the expectation. I did not check your patch carefully, but I think it should be more related with the test environment. It is not strange that the performance is bad under VM environment. Since it is not correctness issue, you can adjust the speed expectation range or skip this failure directly.

[Andreas Dilger]

My concern here is that the DNE changes are slowing down the system somehow, and that is what is causing these different related test failures, since they aren't being seen by any other code.

[Di Wang]

I am not sure about other failures. But this failure happens with only two tiny patches, i.e. no major DNE patches there yet.

http://review.whamcloud.com/#/c/14679/
http://review.whamcloud.com/#/c/14678/7

So this probably not because of DNE patches. And I only saw this one time during all these tests.

[nasf (Inactive)]

The OI scrub logic is relatively independent. Unless some patches introduce or potential cause more background workload (consuming CPU, I/O bandwidth), otherwise, the OI scrub speed should not be affected.

[Di Wang]

Close it for now, since this failure does not happen for almost 2 months. Reopen it when it happen again.

[Bruno Faccini (Inactive)]

+1 for a master patch/review :
https://testing.hpdd.intel.com/test_sets/a709de32-2c8d-11e7-814a-5254006e85c2