Uploaded image for project: 'Lustre'
  1. Lustre
  2. LU-7236

OST connect and disconnect on demand

Details

    • Improvement
    • Resolution: Fixed
    • Minor
    • Lustre 2.12.0
    • None
    • 9223372036854775807

    Description

      currently Lustre client establishes connections to all targets at mount and tries to keep them alive all the time. at big scale this can lead to lots of extra work at mount, failover.

      Attachments

        Issue Links

          is blocked by

          Improvement - An improvement or enhancement to an existing feature or task. LU-10018 MDT as a statfs proxy

          • Minor - Minor loss of function, or other problem where easy workaround is present.
          • Resolved
          is related to

          Bug - A problem which impairs or prevents the functions of the product. LU-14335 sanity test_812b: can't put import into IDLE state after 40 sec, have FULL

          • Minor - Minor loss of function, or other problem where easy workaround is present.
          • Open

          Bug - A problem which impairs or prevents the functions of the product. LU-11518 lock_count is exceeding lru_size

          • Major - Major loss of function.
          • Resolved

          Bug - A problem which impairs or prevents the functions of the product. LU-11951 sanity: test_231a failure, idle disconnect & import generation disagreement

          • Major - Major loss of function.
          • Resolved

          Bug - A problem which impairs or prevents the functions of the product. LU-12328 FLR mirroring on 2.12.1-1 not usable if OST is down

          • Major - Major loss of function.
          • Resolved

          Bug - A problem which impairs or prevents the functions of the product. LU-12799 NULL pointer dereference IP: ptlrpc_disconnect_and_idle_import

          • Major - Major loss of function.
          • Resolved

          Bug - A problem which impairs or prevents the functions of the product. LU-9660 reduce ptlrpcd wakeups on idle system

          • Minor - Minor loss of function, or other problem where easy workaround is present.
          • Resolved

          Bug - A problem which impairs or prevents the functions of the product. LU-11126 replay-single test 89: import.c:1668:ptlrpc_disconnect_idle_interpret()) ASSERTION( imp->imp_state == LUSTRE_IMP_CONNECTING ) failed:

          • Minor - Minor loss of function, or other problem where easy workaround is present.
          • Resolved

          Bug - A problem which impairs or prevents the functions of the product. LU-11128 replay-single test timeout

          • Minor - Minor loss of function, or other problem where easy workaround is present.
          • Resolved

          Bug - A problem which impairs or prevents the functions of the product. LU-11206 recovery-mds-scale test failover_ost fails with 'import is not in FULL state'

          • Minor - Minor loss of function, or other problem where easy workaround is present.
          • Resolved

          Bug - A problem which impairs or prevents the functions of the product. LU-12759 parameter grant_shrink gets reset to 1 after client reconnects

          • Minor - Minor loss of function, or other problem where easy workaround is present.
          • Resolved

          Bug - A problem which impairs or prevents the functions of the product. LU-12857 recovery-mds-scale test_failover_ost fails with “import is not in FULL state”

          • Minor - Minor loss of function, or other problem where easy workaround is present.
          • Resolved

          Bug - A problem which impairs or prevents the functions of the product. LU-12364 idle disconnect shouldn't set IDLE going to reconnect

          • Minor - Minor loss of function, or other problem where easy workaround is present.
          • Closed

          New Feature - A new feature of the product, which has yet to be developed. LU-5703 Quiesce client mountpoints from the server

          • Major - Major loss of function.
          • Open

          Improvement - An improvement or enhancement to an existing feature or task. LU-18015 DNE: client-MDT idle disconnection

          • Minor - Minor loss of function, or other problem where easy workaround is present.
          • Open

          Improvement - An improvement or enhancement to an existing feature or task. LU-9441 Use kernel threads in predictable fashion to confine OS noise

          • Major - Major loss of function.
          • Resolved

          Improvement - An improvement or enhancement to an existing feature or task. LU-9555 "df /path/to/project" should return projid-specific values

          • Minor - Minor loss of function, or other problem where easy workaround is present.
          • Resolved

          Improvement - An improvement or enhancement to an existing feature or task. LU-11405 add a test for idle connection feature

          • Minor - Minor loss of function, or other problem where easy workaround is present.
          • Closed

          Activity

            [LU-7236] OST connect and disconnect on demand
            pjones Peter Jones added a comment -

            Landed for 2.12

            pjones Peter Jones added a comment - Landed for 2.12
            gerrit Gerrit Updater added a comment -

            Oleg Drokin (oleg.drokin@intel.com) merged in patch https://review.whamcloud.com/16682/
            Subject: LU-7236 ptlrpc: idle connections can disconnect
            Project: fs/lustre-release
            Branch: master
            Current Patch Set:
            Commit: 5a6ceb664f07812c351786c1043da71ff5027f8c

            gerrit Gerrit Updater added a comment - Oleg Drokin (oleg.drokin@intel.com) merged in patch https://review.whamcloud.com/16682/ Subject: LU-7236 ptlrpc: idle connections can disconnect Project: fs/lustre-release Branch: master Current Patch Set: Commit: 5a6ceb664f07812c351786c1043da71ff5027f8c
            bzzz Alex Zhuravlev added a comment -

            there was a discussion at LAD. Developers from Fujitsu said that connections are consuming quite amount of memory. Then we recalled that there are other possible cases where less connection can help - like recovery, cluster startup.

            bzzz Alex Zhuravlev added a comment - there was a discussion at LAD. Developers from Fujitsu said that connections are consuming quite amount of memory. Then we recalled that there are other possible cases where less connection can help - like recovery, cluster startup.
            rread Robert Read added a comment -

            Interesting idea. I remember discussing this ages ago when we made the initial connection asynchronous, but don't recall exactly why we didn't do it. Still sounds like a good idea to me.

            rread Robert Read added a comment - Interesting idea. I remember discussing this ages ago when we made the initial connection asynchronous, but don't recall exactly why we didn't do it. Still sounds like a good idea to me.
            gerrit Gerrit Updater added a comment -

            Alex Zhuravlev (alexey.zhuravlev@intel.com) uploaded a new patch: http://review.whamcloud.com/16682
            Subject: LU-7236 ptlrpc: idling connections can disconnect
            Project: fs/lustre-release
            Branch: master
            Current Patch Set: 1
            Commit: 83d9fc0a0a7914174a5373142b183a7bd99be70a

            gerrit Gerrit Updater added a comment - Alex Zhuravlev (alexey.zhuravlev@intel.com) uploaded a new patch: http://review.whamcloud.com/16682 Subject: LU-7236 ptlrpc: idling connections can disconnect Project: fs/lustre-release Branch: master Current Patch Set: 1 Commit: 83d9fc0a0a7914174a5373142b183a7bd99be70a

            People

              bzzz Alex Zhuravlev
              bzzz Alex Zhuravlev
              Votes:
              0 Vote for this issue
              Watchers:
              11 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: