Details

    • Bug
    • Resolution: Fixed
    • Critical
    • Lustre 2.7.0
    • Lustre 2.7.0
    • 3
    • 16091

    Description

      This issue was created by maloo for Nathaniel Clark <nathaniel.l.clark@intel.com>

      This issue relates to the following test suite run:
      https://testing.hpdd.intel.com/test_sets/8e09d60c-48fc-11e4-be93-5254006e85c2
      https://testing.hpdd.intel.com/test_sets/8384ce00-5179-11e4-b4cf-5254006e85c2

      The sub-test test_18a failed with the following error:

      (6.1) Expect 1 fixed on mds1, but got: 2
      

      Please provide additional information about the failure here.

      Info required for matching: sanity-lfsck 18a

      Attachments

        Activity

          [LU-5731] sanity-lfsck test_18a:

          Oleg Drokin (oleg.drokin@intel.com) merged in patch http://review.whamcloud.com/12359/
          Subject: LU-5731 osp: flush async updates for osp_sync
          Project: fs/lustre-release
          Branch: master
          Current Patch Set:
          Commit: 138a8d14b48951b32a25d2ceada583bfdba494ce

          gerrit Gerrit Updater added a comment - Oleg Drokin (oleg.drokin@intel.com) merged in patch http://review.whamcloud.com/12359/ Subject: LU-5731 osp: flush async updates for osp_sync Project: fs/lustre-release Branch: master Current Patch Set: Commit: 138a8d14b48951b32a25d2ceada583bfdba494ce
          yujian Jian Yu added a comment - One more instance on master branch: https://testing.hpdd.intel.com/test_sets/71f79e8a-59e7-11e4-aa32-5254006e85c2
          yong.fan nasf (Inactive) added a comment - Here is the patch: http://review.whamcloud.com/#/c/12359/

          sanity-lfsck test_18a assume that before the test start, there were no orphan OST-objects in the system. It is the test_18a itself to generate orphan OST-object during the test. Unfortunately, there were some orphan OST-object left from other former tests. So more orphan OST-objects have been found and repaired by test_18a. What we need to do is NOT to adjust the test scripts, but found out why there are orphan OST-objects before the test_18a.

          After some investigation, I found that in sanity test_14, the async updates for LFSCK have not been handled completely when the LFSCK finished, although before the LFSCK exit, it called dt_sync(), but current dt_sync() ignores the async updates that are handled by background ptlrpcd threads directly. Then those async LFSCK updates generated some orphan OST-objects, and made the test environment dirty.

          yong.fan nasf (Inactive) added a comment - sanity-lfsck test_18a assume that before the test start, there were no orphan OST-objects in the system. It is the test_18a itself to generate orphan OST-object during the test. Unfortunately, there were some orphan OST-object left from other former tests. So more orphan OST-objects have been found and repaired by test_18a. What we need to do is NOT to adjust the test scripts, but found out why there are orphan OST-objects before the test_18a. After some investigation, I found that in sanity test_14, the async updates for LFSCK have not been handled completely when the LFSCK finished, although before the LFSCK exit, it called dt_sync(), but current dt_sync() ignores the async updates that are handled by background ptlrpcd threads directly. Then those async LFSCK updates generated some orphan OST-objects, and made the test environment dirty.

          People

            yong.fan nasf (Inactive)
            maloo Maloo
            Votes:
            0 Vote for this issue
            Watchers:
            6 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved: