Retry directory destruction in ResetDbpath

XMLWordPrintableJSON

    • Type: Bug
    • Resolution: Fixed
    • Priority: Major - P3
    • 5.2.0
    • Affects Version/s: None
    • Component/s: None
    • None
    • Fully Compatible
    • ALL
    • Security 2021-11-15
    • 40
    • None
    • 3
    • None
    • None
    • None
    • None
    • None
    • None

      The ResetDbpath shell helper is used to destroy old database paths created by the integration tests.

      Unfortunately, sometimes calls to boost::filesystem::create_directory appear to fail spuriously on Windows. This is likely due to file handles remaining active in the kernel, either because of some background job, or because cleanup takes longer than expected. We had previously experienced faults while performing boost::filesystem::remove_all, but resolved that by retrying errors on Windows in SERVER-44553.

      While there are likely more comprehensive solutions, they would likely be more expensive to implement and maintain. We should investigate whether a simple retry would meet our needs.

            Assignee:
            Mark Benvenuto
            Reporter:
            Spencer Jackson
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

              Created:
              Updated:
              Resolved: