Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-38402

closeAll_with_background_ops_fails_safely relies on an incorrect assumption

    XMLWordPrintableJSON

Details

    • Icon: Bug Bug
    • Resolution: Won't Fix
    • Icon: Minor - P4 Minor - P4
    • None
    • 4.0.0, 4.1.1
    • Storage
    • None
    • Storage Execution
    • ALL
    • v4.0
    • 21

    Description

      The test attempts to assert that restarting the catalog will correctly restore any state in the event the operation fails. To do this it starts a background index build (which causes closing the catalog to fail), then calls restartCatalog. However, the test calls restartCatalog after it observes the index build in currentOp, which it turns out is not sufficient. The server code will first register the command in CurOp before marking it as a BackgroundOperation. Thus it's possible for the restartCatalog to succeed as it can come in before the index build marks itself as a BackgroundOperation.

      Attachments

        Activity

          People

            backlog-server-execution Backlog - Storage Execution Team
            daniel.gottlieb@mongodb.com Daniel Gottlieb (Inactive)
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved: