Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-55571

move_chunk_critical_section_non_internal_client_abort.js fails because moveChunk cmd might get interrupted twice in some suites. (like sharding_csrs_continuous_config_stepdown)

    • Fully Compatible
    • ALL
    • Sharding 2021-06-14, Sharding 2021-06-28
    • 130

      In move_chunk_critical_section_non_internal_client_abort.js moveChunk cmd will fail with an Interrupted error code because of collMod (this is expected). In the suite, sharding_csrs_continuous_config_stepdown, a csrs node will step down while the test is running. If this happens after the first interruption while the _configsvrMoveChunk cmd is being executed, it will fail with a InterruptedDueToReplStateChange error. This is a retriable error. If this retriable error happens after the first Interruption due to collMod, collMod won't run again, so the retry of _configsvrMoveChunk cmd will not be interrupted and will succeed. This might cause the test to fail with a: command worked when it should have failed exception.

            Assignee:
            haley.connelly@mongodb.com Haley Connelly
            Reporter:
            kshitij.gupta@mongodb.com Kshitij Gupta
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

              Created:
              Updated:
              Resolved: