Do not wait for range deletion thread on stepdown

XMLWordPrintableJSON

    • Type: Bug
    • Resolution: Fixed
    • Priority: Major - P3
    • 6.2.0-rc0
    • Affects Version/s: 6.2.0-rc0
    • Component/s: None
    • None
    • Fully Compatible
    • ALL
    • Hide

      Apply the provided patch on top of r6.2.0-alpha-1295-gab34541c92e and run:

      resmoke.py --suite=sharding jstests/sharding/range_deletion_hang_stepdown.js
      Show
      Apply the provided patch on top of r6.2.0-alpha-1295-gab34541c92e and run: resmoke.py --suite=sharding jstests/sharding/range_deletion_hang_stepdown.js
    • Sharding EMEA 2022-10-31, Sharding EMEA 2022-11-14
    • 55
    • None
    • 3
    • None
    • None
    • None
    • None
    • None
    • None
    • None

      On stepdown the RangeDeleterService destruct the ReadyRangeDeletionsProcessor that in turns interrupt and join its internal thread.

      In order not to slowdown the stepdown process we should instead

      • Interrupt the internal thread onStepDown()
      • Join the internal thread onStepUpBegin()
      • Start the internal thread onStepUpComplete

      The proposed approach will still guarantee that all the previous range deletions will be drained strictly before the node can transition to PRIMARY again.

        1. SERVER-70964-reproducible.patch
          3 kB

              Assignee:
              Tommaso Tocci
              Reporter:
              Tommaso Tocci
              Votes:
              0 Vote for this issue
              Watchers:
              3 Start watching this issue

                Created:
                Updated:
                Resolved: