[LU-10199] sanity test_56xb: hard links individually migrated (12 != 11)

• Type: Bug
• Resolution: Fixed
• Priority: Major
• Fix Version/s: Lustre 2.11.0, Lustre 2.10.4
• Affects Version/s: Lustre 2.11.0
• Severity: 3
• Assignee: Steve Guminski (Inactive)
• Reporter: Maloo

Description

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

This issue relates to the following test suite run: https://testing.whamcloud.com/test_sets/740cdf26-c31d-11e7-88ab-52540065bddc.

The sub-test test_56xb failed with the following error:

hard links individually migrated (12 != 11)

strongly suspect this is due to the very recently landed "LU-6051 utils: allow lfs_migrate to handle hard links"
test 56xb is brand new, only introduced in this patch.
I'm wondering if this test has ever been passed on any SLES version.

Info required for matching: sanity 56xb

Comments

[Gerrit Updater]

Steve Guminski ([e-mail]) uploaded a new patch: https://review.whamcloud.com/30616
Subject: LU-10199 tests: Re-enable sanity test 56xb
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 14c03c1e43f62fa79fc38288834ebc0a5df40cf2

[Peter Jones]

Landed for 2.11

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch https://review.whamcloud.com/30078/
Subject: LU-10199 utils: strip // from fid2path
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 116b73d69bf2462fa9e9b8b9f629794bbda90ee5

[Gerrit Updater]

Ben Evans ([e-mail]) uploaded a new patch: https://review.whamcloud.com/30078
Subject: LU-10199 utils: strip // from fid2path
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 5a87fb56b121fdcd623cb9298e53903ee5719ff1

[Andreas Dilger]

It would be good to have some examples of how this is broken, since I can't imagine how "lfs fid2path" is generating a non-absolute pathnames.

[Gerrit Updater]

Steve Guminski ([e-mail]) uploaded a new patch: https://review.whamcloud.com/30013
Subject: LU-10199 lfs: Fix file name comparison in lfs_migrate
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 3a7fd98d46729b0d430b6fe30c55cba275574303

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch https://review.whamcloud.com/29954/
Subject: LU-10199 utils: disable sanity test 56xb
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 3ac484707aecb97ab8aa4716110205108b0535ea

[Steve Guminski (Inactive)]

Yes, that's possible.  My local CentOS test system was upgraded during that period.

[Bob Glossman (Inactive)]

Here is an instance of it failing on el7 too:
https://testing.hpdd.intel.com/test_sets/216f1a9c-c361-11e7-a943-52540065bddc

Is it possible that it originally passed test on el7.3 and is now failing on el7.4?

[Steve Guminski (Inactive)]

I just reran the sanity test on my local RHEL system, and it fails there, too.  So I'm not sure what has changed between the time that the patch was originally submitted and now that would cause this change in behavior.  I'll submit a new patch that disables the test instead of reverting.

[Andreas Dilger]

Do we have any idea why this is failing? I'd prefer a patch that fixes a problem vs. reverting it. Next best would be a patch to skip just the failing test for SLES, so that we keep the patch landed and testing under RHEL, and give us some time to fix the SLES issue.