[LU-5624] sanity-lfsck test_9b: Got speed 63, expected more than 64

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.7.0
• Affects Version/s: Lustre 2.7.0
• Labels: performance zfs
• Severity: 3
• Assignee: nasf (Inactive)
• Reporter: Maloo

Description

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

This issue relates to the following test suite runs: https://testing.hpdd.intel.com/test_sets/c9f806be-3b57-11e4-ad5c-5254006e85c2
https://testing.hpdd.intel.com/test_sets/a59559ac-277f-11e4-ad6f-5254006e85c2
https://testing.hpdd.intel.com/test_sets/908bc4e4-2e53-11e4-9cc5-5254006e85c2

The sub-test test_9b failed with the following error:

(9) Got speed 63, expected more than 64

Please provide additional information about the failure here.

Info required for matching: sanity-lfsck 9b

Issue Links

• is related to: LU-8877 sanity-lfsck test_9a error

Comments

[nasf (Inactive)]

The patch has been landed to Lustre-2.7

[nasf (Inactive)]

I made a patch that ignore the bad performance for ZFS-based backend since we do not know how slow the ZFS will be, although we already give some low margin. In fact, we have already done that in LU-4934, but related patch did not work as expected.

http://review.whamcloud.com/12322

[Andreas Dilger]

Instead of disabling the test for ZFS, it would be better to give it some margin for running slower without failing. That will avoid potential regressions in this code, while allowing the test to still pass.

[nasf (Inactive)]

It is not surprise that ZFS backend run slowly in VM environment. I am thinking that we should disable speed related tests for ZFS backend until we really resolved ZFS performance.