[LU-10047] sanity test_102d: mkdir -i 2 -c2 /mnt/lustre/d102d.sanity failed

• Status: Resolved
• Project: Lustre
• Fix Version/s: Lustre 2.11.0, Lustre 2.10.2
• Type: Bug
• Priority: Minor
• Reporter: Maloo
• Assignee: Bob Glossman (Inactive)
• Resolution: Fixed
• Severity: 3

Description

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

This issue relates to the following test suite run: https://testing.hpdd.intel.com/test_sets/b502a26a-a4a8-11e7-b786-5254006e85c2.

The sub-test test_102d failed with the following error:

mkdir -i 2 -c2 /mnt/lustre/d102d.sanity failed

similar fail also seen on other test_102 subtests
I wonder if this is related to LU-10024.

Info required for matching: sanity 102d

Issue Links

• is related to: LU-9476 sanity tests 102d, f, j skipped looki...
• is related to: LU-10108 sanity/102d fails due to already-crea...

Comments

[Bob Glossman (Inactive)]

seen again on master:
https://testing.hpdd.intel.com/test_sets/e974e1f4-a53c-11e7-bb19-5254006e85c2

The fact that it happened more than once is suspicious.
The ones I've seen both happened on el6.9, not el7.
Don't know if this is a cause or not.

No fails like this on other branches, only master.

[Bob Glossman (Inactive)]

When run on el7 the failing tests are skipped with:

lustre-aware tar is not installed

This probably explains why the fail is only seen on el6, not el7.
I strongly suspect it will fail anytime it's run, but it's not run on el7 due to skip test logic.

[John Hammond]

This was introduced by the following change from https://review.whamcloud.com/26212 LU-3308 tests: fix sanity/sanityn test_mkdir() usage.

@@ -7139,12 +7103,12 @@ test_102d() {
        [ -z "$TAR" ] && skip_env "lustre-aware tar is not installed" && return
        [[ $OSTCOUNT -lt 2 ]] && skip_env "needs >= 2 OSTs" && return
        setup_test102
-       test_mkdir -p $DIR/d102d
-       $TAR xf $TMP/f102.tar -C $DIR/d102d --xattrs
-       cd $DIR/d102d/$tdir
+       test_mkdir $DIR/$tdir
+       $TAR xf $TMP/$tfile.tar -C $DIR/$tdir --xattrs
+       cd $DIR/$tdir/$tdir
        compare_stripe_info1
 }
-run_test 102d "tar restore stripe info from tarfile,not keep osts ==========="
+run_test 102d "tar restore stripe info from tarfile,not keep osts"
 
 test_102f() {
        [ $PARALLEL == "yes" ] && skip "skip parallel run" && return

We need the -p option when we call test_mkdir().

[Bob Glossman (Inactive)]

seems like that mod for LU-3308 had a lot of bad side effects. First LU-10024, now LU-10047. wonder if there are more lurking we haven't noticed yet.

[Gerrit Updater]

Bob Glossman ([e-mail]) uploaded a new patch: https://review.whamcloud.com/29275
Subject: LU-10047 test: add -p options back into ssnity tests
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 909a0e0dae2a4e2322aef38dbfc4333706c32746

[Bob Glossman (Inactive)]

at least one ok test run with the patch shows that it does in fact work.
in https://testing.hpdd.intel.com/test_sessions/b03f80f9-f232-4533-b4ef-cf524a78d332 sanity tests 102d and 102j are run, not skipped, and don't fail.

[Gerrit Updater]

Bob Glossman ([e-mail]) uploaded a new patch: https://review.whamcloud.com/29288
Subject: LU-10047 tests: stop skipping test_102 subtests
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: a733b59d139e05623744e0dd9f9a489ca72e4c44

[Bob Glossman (Inactive)]

Possible better fix.
Eliminates unneeded version checks and test_mkdir() calls.

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch https://review.whamcloud.com/29288/
Subject: LU-10047 tests: stop skipping test_102 subtests
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 036641a1e1f6d5d010873267b669528890f19fe4

[Gerrit Updater]

Bob Glossman ([e-mail]) uploaded a new patch: https://review.whamcloud.com/29823
Subject: LU-10047 tests: stop skipping test_102 subtests
Project: fs/lustre-release
Branch: b2_10
Current Patch Set: 1
Commit: efbcd788905ae9664a89b70afe378e7b037cacf1

[Gerrit Updater]

John L. Hammond ([e-mail]) merged in patch https://review.whamcloud.com/29823/
Subject: LU-10047 tests: stop skipping test_102 subtests
Project: fs/lustre-release
Branch: b2_10
Current Patch Set:
Commit: f74cafb93d385e5a425424f7bad4e9b36cc3aad4

[Gerrit Updater]

Andreas Dilger ([e-mail]) uploaded a new patch: https://review.whamcloud.com/30251
Subject: LU-10047 backup: add --xattrs-include="trusted.*"
Project: doc/manual
Branch: master
Current Patch Set: 1
Commit: 13355a3bf0f93da891ad5df209f7a77b2bee03eb

[Gerrit Updater]

Joseph Gmitter ([e-mail]) merged in patch https://review.whamcloud.com/30251/
Subject: LU-10047 backup: add --xattrs-include="trusted.*"
Project: doc/manual
Branch: master
Current Patch Set:
Commit: 9a4924650689205147eec94a52e07093483a1f7a