[LU-3812] Interop 2.4.0<->2.5 failure on test suite sanity-scrub test_10b: Expect 'paused', but got 'stopped'

• Type: Bug
• Resolution: Fixed
• Priority: Critical
• Fix Version/s: Lustre 2.4.2
• Affects Version/s: Lustre 2.4.0, Lustre 2.4.1
• Environment: server: lustre-master tag-2.4.91 build #1627; client: 2.4.0
• Severity: 3
• Assignee: Jian Yu
• Reporter: Maloo

Description

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

This issue relates to the following test suite run: http://maloo.whamcloud.com/test_sets/9a665dec-09ee-11e3-ad93-52540035b04c.

The sub-test test_10b failed with the following error:

(9) Expect 'paused', but got 'stopped'

Info required for matching: sanity-scrub 10b

Comments

[Peter Jones]

Landed for 2.4.2

[Jian Yu]

Patch for Lustre b2_4 branch: http://review.whamcloud.com/7529

[Jian Yu]

The patch on master branch is:

commit f5491211ed7251a45e4e6916539cd2a2ab5e6038
Author: Fan Yong <[e-mail]>
Date:   Sun May 19 12:26:37 2013 +0800

    LU-2915 lfsck: LFSCK 1.5 technical debts (2)
    
    This patch resolves some LFSCK 1.5 technical debts, including:
    1) LFSCK does NOT hold spin_lock when call dt iteration API(s).
    2) Remove the hack of using dt_it_ops::put() method for events
       notification to low layer otable-based iteration.
    
    Test-Parameters: testlist=sanity-scrub,sanity-lfsck
    Signed-off-by: Fan Yong <[e-mail]>
    Change-Id: Ie16161370ac218860a74b23d853c4c378adce4ec
    Reviewed-on: http://review.whamcloud.com/6343
    Tested-by: Hudson
    Tested-by: Maloo <[e-mail]>
    Reviewed-by: Alex Zhuravlev <[e-mail]>
    Reviewed-by: Andreas Dilger <[e-mail]>

diff --git a/lustre/tests/sanity-scrub.sh b/lustre/tests/sanity-scrub.sh
<~snip~>
+# test_10b is obsolete, it will be coverded by related sanity-lfsck tests.
 test_10b() {
        scrub_prep 0
        mds_backup_restore || error "(1) Fail to backup/restore!"
@@ -742,7 +743,7 @@ test_10b() {
        FLAGS=$($SHOW_SCRUB | awk '/^flags/ { print $2 }')
        [ -z "$FLAGS" ] || error "(14) Expect empty flags, but got '$FLAGS'"
 }
-run_test 10b "non-stopped OI scrub should auto restarts after MDS remount (2)"
+#run_test 10b "non-stopped OI scrub should auto restarts after MDS remount (2)"

As per the suggestion from Nasf, we only need add Lustre version check code into sanity-scrub test 10b on Lustre b2_4 branch to resolve the interop failure. I'll upload the patch.

[Jian Yu]

Hi nasf,

Do you know which patch on master branch fixed this issue?

[Jodi Levi (Inactive)]

Oleg believes there is a fix in 2.5 for this that just needs to be ported to 2.4.
Could you please take a look at this?
Thank you!