Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-19092

ConnectionPool's per-host list ends up with deleted connections

XMLWordPrintableJSON

    • Type: Icon: Bug Bug
    • Resolution: Won't Fix
    • Priority: Icon: Critical - P2 Critical - P2
    • None
    • Affects Version/s: None
    • Component/s: Networking
    • ALL
    • Hide

      It is non-deterministic, but seems to happen more frequently if the cleanup interval is lowered:

      const Seconds kCleanUpInterval(2);  // Note: Must be larger than kMaxConnectionAge below)
const Seconds kMaxConnectionAge(1);

      Only happens in the optimized build. Command line to run the test:

      buildscripts\resmoke.py jstests/concurrency/fsm_all_sharded_replication.js > output.log
      Show
      It is non-deterministic, but seems to happen more frequently if the cleanup interval is lowered: const Seconds kCleanUpInterval(2); // Note: Must be larger than kMaxConnectionAge below) const Seconds kMaxConnectionAge(1); Only happens in the optimized build. Command line to run the test: buildscripts\resmoke.py jstests/concurrency/fsm_all_sharded_replication.js > output.log
    • Sharding A (10/09/15)
    • 0
    • None
    • 0
    • None
    • None
    • None
    • None
    • None
    • None

      The ConnectionPool class is used by both replication and sharding. The sharding FSM tests when run sufficiently long encounter a bug where a deleted entry appears in the per-client pool.

            Assignee:
            schwerin@mongodb.com Andy Schwerin
            Reporter:
            kaloian.manassiev@mongodb.com Kaloian Manassiev
            Votes:
            0 Vote for this issue
            Watchers:
            12 Start watching this issue

              Created:
              Updated:
              Resolved: