Uploaded image for project: 'Lustre'
  1. Lustre
  2. LU-15683

sanity-flr.sh test_50c: proper initialize fid in OFD objects fallocate handler

Details

    • Bug
    • Resolution: Fixed
    • Major
    • Lustre 2.16.0
    • None
    • None
    • 3
    • 9223372036854775807

    Description

      the filter_fid is not initialized before storing into the XATTR value.

      Attachments

        Issue Links

          is duplicated by

          Bug - A problem which impairs or prevents the functions of the product. LU-15333 lfsck reports "XATTR trusted.fid: the parent FID is invalid"

          • Major - Major loss of function.
          • Resolved
          is related to

          Bug - A problem which impairs or prevents the functions of the product. LU-15519 fallocate does not increase projid quota usage

          • Major - Major loss of function.
          • Resolved

          Activity

            [LU-15683] sanity-flr.sh test_50c: proper initialize fid in OFD objects fallocate handler
            pjones Peter Jones added a comment -

            Landed for 2.16

            pjones Peter Jones added a comment - Landed for 2.16
            gerrit Gerrit Updater added a comment -

            "Oleg Drokin <green@whamcloud.com>" merged in patch https://review.whamcloud.com/46919/
            Subject: LU-15683 ofd: proper initialize filter_fid in ofd fallocate
            Project: fs/lustre-release
            Branch: master
            Current Patch Set:
            Commit: 755fdc6678836652302cea51f838e4894f2a7efc

            gerrit Gerrit Updater added a comment - "Oleg Drokin <green@whamcloud.com>" merged in patch https://review.whamcloud.com/46919/ Subject: LU-15683 ofd: proper initialize filter_fid in ofd fallocate Project: fs/lustre-release Branch: master Current Patch Set: Commit: 755fdc6678836652302cea51f838e4894f2a7efc
            bobijam Zhenyu Xu added a comment - - edited

            yes, mirror resync uses fallocate to punch hole in destination mirror. llapi_mirror_resync_many()->llapi_mirror_punch()->llapi_hole_punch()->fallocate() 

            bobijam Zhenyu Xu added a comment - - edited yes, mirror resync uses fallocate to punch hole in destination mirror. llapi_mirror_resync_many()->llapi_mirror_punch()->llapi_hole_punch()->fallocate() 
            adilger Andreas Dilger added a comment - - edited

            Maybe I'm missing something, but I don't see sanity-flr test_50b calling fallocate? Is it because "lfs mirror resync" is using PUNCH when resyncing the file with holes?

            adilger Andreas Dilger added a comment - - edited Maybe I'm missing something, but I don't see sanity-flr test_50b calling fallocate? Is it because "lfs mirror resync" is using PUNCH when resyncing the file with holes?
            bobijam Zhenyu Xu added a comment -

            Oh, I see that patch is not landed yet, so maybe it is also the fix?

            I think this is another issue I find during investigating https://testing.whamcloud.com/test_sessions/8e4a262c-d321-4842-b992-b6eb5083407a sanity-flr.sh test_50c failure, the OFD object create its layout XATTR_NAME_FID with random layout version, and that the supposed correctly write request got rejected since the layout version passed from the client does not exceeds the random big layout version number.

            bobijam Zhenyu Xu added a comment - Oh, I see that patch is not landed yet, so maybe it is also the fix? I think this is another issue I find during investigating https://testing.whamcloud.com/test_sessions/8e4a262c-d321-4842-b992-b6eb5083407a sanity-flr.sh test_50c failure, the OFD object create its layout XATTR_NAME_FID with random layout version, and that the supposed correctly write request got rejected since the layout version passed from the client does not exceeds the random big layout version number.
            arshad512 Arshad Hussain added a comment -

            >Oh, I see that patch is not landed yet, so maybe it is also the fix?

            Andreas, we are also missing passing the OBD_MD_FLPROJID flag from vvp,osc etc... Which is taken care by LU-15519

            arshad512 Arshad Hussain added a comment - >Oh, I see that patch is not landed yet, so maybe it is also the fix? Andreas, we are also missing passing the OBD_MD_FLPROJID flag from vvp,osc etc... Which is taken care by LU-15519
            gerrit Gerrit Updater added a comment -

            "Bobi Jam <bobijam@hotmail.com>" uploaded a new patch: https://review.whamcloud.com/46919
            Subject: LU-15683 ofd: proper initialize filter_fid in ofd fallocate
            Project: fs/lustre-release
            Branch: master
            Current Patch Set: 1
            Commit: 54c8c7a9a83f4175cefebdbdce07a14211fd2e1f

            gerrit Gerrit Updater added a comment - "Bobi Jam <bobijam@hotmail.com>" uploaded a new patch: https://review.whamcloud.com/46919 Subject: LU-15683 ofd: proper initialize filter_fid in ofd fallocate Project: fs/lustre-release Branch: master Current Patch Set: 1 Commit: 54c8c7a9a83f4175cefebdbdce07a14211fd2e1f
            adilger Andreas Dilger added a comment -

            Oh, I see that patch is not landed yet, so maybe it is also the fix?

            adilger Andreas Dilger added a comment - Oh, I see that patch is not landed yet, so maybe it is also the fix?
            adilger Andreas Dilger added a comment -

            Is this issue the same as LU-15519, or is it caused by the landing of "LU-15519 quota: fallocate does not increase projectid usage"?

            adilger Andreas Dilger added a comment - Is this issue the same as LU-15519 , or is it caused by the landing of " LU-15519 quota: fallocate does not increase projectid usage "?

            People

              bobijam Zhenyu Xu
              bobijam Zhenyu Xu
              Votes:
              0 Vote for this issue
              Watchers:
              5 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: