[LU-8326] sanity-quota test_21 test failed to respond and timed out

• Type: Bug
• Resolution: Duplicate
• Priority: Blocker
• Fix Version/s: Lustre 2.10.0
• Affects Version/s: Lustre 2.9.0, Lustre 2.10.0
• Labels: zfs
• Severity: 3
• Assignee: Niu Yawei (Inactive)
• Reporter: Maloo

Description

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

Please provide additional information about the failure here.

This issue relates to the following test suite run: https://testing.hpdd.intel.com/test_sets/648de7b4-3a40-11e6-bbf5-5254006e85c2.

Issue Links

• is related to: LU-4448 Test failure timeout on sanity-quota test_7c: client sysrq and reboot
• is related to: LU-2435 inode accounting in osd-zfs is racy

Comments

[Niu Yawei (Inactive)]

dup of LU-2435.

[Peter Jones]

Has this been hit since LU-2435 landed on March 30th? If not, then can we mark it resolved as a dupe?

[Shuichi Ihara (Inactive)]

+1 on master
https://testing.hpdd.intel.com/test_sessions/e930b05a-d1c3-4ab9-ab8d-fa8e222f9d54

[Henri Doreau (Inactive)]

+1 on master:
https://testing.hpdd.intel.com/test_sets/8079c25a-085c-11e7-a090-5254006e85c2

[Emoly Liu]

+1 on master:
https://testing.hpdd.intel.com/test_sets/4772f1a6-07d6-11e7-9ad1-5254006e85c2

[Andreas Dilger]

As discussed today, we will try to get patch https://review.whamcloud.com/15294 "LU-2435 osd-zfs: use zfs native dnode accounting" landed, since that is actually fixing the problem rather than skipping the test. It should be very close to landing.

[James Nunez (Inactive)]

Uploaded a patch, https://review.whamcloud.com/25919, to add sanity-quota test 21 to the ALWAYS_EXCEPT list for ZFS testing in case we decide to stop running the test .... temporarily.

[Gerrit Updater]

James Nunez ([e-mail]) uploaded a new patch: https://review.whamcloud.com/25919
Subject: LU-8326 test: skip sanity-quota 21 for ZFS
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 6af31b54dbf5783b067c24d0c712b1127b8f1ee1

[Alex Zhuravlev]

LU-2435 has been around and ready for months.. let's not disable the tests, but land the fix instead?

[Joseph Gmitter (Inactive)]

Per discussion with Niu, he believes that this is caused by slow zfs performance and when all the related ZFS patches are deployed to the test system that the problem should be resolved. Niu recommends disabling the test if the zfs patches can not be deployed to the test system quickly.

Niu, do you have any more detail you would like to add?