Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-40184

Crash in server after bulk write error

    • Type: Icon: Bug Bug
    • Resolution: Duplicate
    • Priority: Icon: Major - P3 Major - P3
    • None
    • Affects Version/s: 4.1.9
    • Component/s: None
    • None
    • Sharding
    • ALL
    • Hide

      I am creating the sharded cluster using mlaunch:
      {{ VERSION=4.1.9}}
      {{ MONGO_PATH=/usr/local/m/versions/${VERSION}/bin}}mlaunch --binarypath ${MONGO_PATH} --replicaset --name "repl0" --sharded a1 a2 a3 --mongos 3 --dir ~/dbData --setParameter enableTestCommands=1 -vvvv

      So far, I can only reproduce this on my Macbook by running the transactions test runner for the Java driver.

      Show
      I am creating the sharded cluster using mlaunch: {{ VERSION=4.1.9}} {{ MONGO_PATH=/usr/local/m/versions/${VERSION}/bin}} mlaunch --binarypath ${MONGO_PATH} --replicaset --name "repl0" --sharded a1 a2 a3 --mongos 3 --dir ~/dbData --setParameter enableTestCommands=1 -vvvv So far, I can only reproduce this on my Macbook by running the transactions test runner for the Java driver.

      While testing my implementation for sharded transaction recovery token using the transactions tests in the Java driver, the server crashes on one of the tests.

      My Macbook unexpectedly shut down, and after restarting, the crash occurs on a transactions test.

      When running the transactions test, the log shows the expected `WriteConflict` error, followed immediately by an abort. I am attaching the log file that contains the abort and stack trace.

       

            Assignee:
            backlog-server-sharding [DO NOT USE] Backlog - Sharding Team
            Reporter:
            john.stewart@mongodb.com John Stewart (Inactive)
            Votes:
            0 Vote for this issue
            Watchers:
            6 Start watching this issue

              Created:
              Updated:
              Resolved: