[LU-17778] conf-sanity test_76d: 'llite.*.xattr_cache != 0 on client /mnt/lustre'

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.17.0
• Severity: 3
• Assignee: Andreas Dilger
• Reporter: Maloo

Description

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

This issue relates to the following test suite run: https://testing.whamcloud.com/test_sets/e0acabcb-3a96-4537-a095-80228445c727

test_76d failed with the following error:

llite.*.xattr_cache != 0 on client /mnt/lustre

Test session details:
clients: https://build.whamcloud.com/job/lustre-reviews/104338 - 5.14.0-362.18.1.el9_3.x86_64
servers: https://build.whamcloud.com/job/lustre-reviews/104338 - 5.14.0-362.18.1_lustre.el9.x86_64

VVVVVVV DO NOT REMOVE LINES BELOW, Added by Maloo for auto-association VVVVVVV
conf-sanity test_76d - llite.*.xattr_cache != 0 on client /mnt/lustre

Issue Links

• is related to: LU-9399 lctl conf_param and lctl set_param -P are broken for llite parameters

Comments

[Peter Jones]

Merged for 2.17

[Gerrit Updater]

"Oleg Drokin <[e-mail]>" merged in patch https://review.whamcloud.com/c/fs/lustre-release/+/58215/
Subject: LU-17778 tests: fix conf-sanity/76d issues
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 129ac1bbd779249f681270616737d1dda993c83b

[Gerrit Updater]

"Andreas Dilger <[e-mail]>" uploaded a new patch: https://review.whamcloud.com/c/fs/lustre-release/+/58215
Subject: LU-17778 tests: fix conf-sanity/76d issues
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: f03cc7ce971b8aa69c91ee820f946c0adb2ed26c

[Andreas Dilger]

This is likely a bug in the way that the test is written and/or how it is running in the VMs. The test is "changing" the value of the llite.*.xattr_cache parameter from 0->1 and 1->0, but since this is traveling via the MGS persistent parameters, there may occasionally be some interval after the client filesystem is mounted when the parameter was not yet updated:

        mount_client $MOUNT2 || error "mount $MOUNT2 failed"
        [ $($LCTL get_param -n llite.*.xattr_cache) -eq $new ] ||
                error "$xattr_cache != $new on client $MOUNT"

It probably makes sense to change these checks to also wait to confirm that the parameter has updated:

        mount_client $MOUNT2 || error "mount $MOUNT2 failed"
        wait_update $HOSTNAME "$cmd" "$new" ||
                error "$xattr_cache != $new on client $MOUNT"

Also, the last part of the subtest that is checking the parameter on the second mountpoint is ineffective, since the "cmd" execution is always getting the first parameter (alphabetically), so it may be ignoring the second parameter in 50% of test runs.

[Andreas Dilger]

+1 on master: https://testing.whamcloud.com/test_sets/06cf346f-4f10-4696-998f-b5d7529c4dbc