[LU-8600] sanity test_399:fake write is slower

• Type: Bug
• Resolution: Fixed
• Priority: Blocker
• Fix Version/s: Lustre 2.9.0
• Affects Version/s: Lustre 2.9.0
• Environment: Lustre version: 2.8.57.39.gbc59a76; master
• Severity: 3
• Assignee: Jinshan Xiong (Inactive)
• Reporter: Maloo

Description

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

This issue relates to the following test suite run: https://testing.hpdd.intel.com/test_sets/611ece5e-75c8-11e6-b058-5254006e85c2.

The sub-test test_399 failed with the following error:

fake write is slower

Test logs:

== sanity test 399: fake write should not be slower than normal write ================================ 02:54:34 (1473328474)
debug=0
1000+0 records in
1000+0 records out
1048576000 bytes (1.0 GB) copied, 543.515 s, 1.9 MB/s
CMD: onyx-50vm4 /usr/sbin/lctl set_param fail_loc=0x238
fail_loc=0x238
1000+0 records in
1000+0 records out
1048576000 bytes (1.0 GB) copied, 632.552 s, 1.7 MB/s
/mnt/lustre/f399.sanity has type file OK
/mnt/lustre/f399.sanity has size 1048576000 OK
CMD: onyx-50vm4 /usr/sbin/lctl set_param fail_loc=0
fail_loc=0
fake write 632.563274817 vs. normal write 543.527018226 in seconds
 sanity test_399: @@@@@@ FAIL: fake write is slower 

Issue Links

• is related to: LU-7655 To drop write data on OST to help performance benchmark

Comments

[Gerrit Updater]

Andreas Dilger ([e-mail]) merged in patch http://review.whamcloud.com/22427/
Subject: LU-8600 tests: ignore error if running in vm for sanity 399
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 2fd108171fb08563d2e0ddc09789620d402b0014

[Gerrit Updater]

Jinshan Xiong ([e-mail]) uploaded a new patch: http://review.whamcloud.com/22427
Subject: LU-8600 tests: ignore error if running in vm for sanity 399
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 27d3c8ceb758c7b9c2a83834e26cd2b6652c27dc

[Andreas Dilger]

Sadly, as I suspected in my last review of http://review.whamcloud.com/5164 this is causing test failures in the VM environment.

Jinshan, can you please submit a patch to check for VM and use error_ignore in case the performance is worse. We still want to run the test, to make sure that the functionality is working.