Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-18799

Reduce flakiness of max_time_ms.js

    XMLWordPrintableJSON

Details

    • Icon: Improvement Improvement
    • Resolution: Done
    • Icon: Major - P3 Major - P3
    • 3.1.4
    • 3.1.3
    • Testing Infrastructure
    • None
    • Fully Compatible
    • Quint Iteration 4
    • 15

    Description

      max_time_ms.js has been failing periodically expecting to time out on a getMore. The theory is that this happens because the timeout is measured using an approximation of the current time, which can become inaccurate under high CPU load. We should only allow the core suites to use as many jobs as there are cores available, and increase the delta between the sleep in the $where clause of the query, and the timeout specified. These will hopefully both reduce the frequency with which this failure occurs.

      We would need to improve our timing logic in the server to make sure the test never sporadically fails. That work should happen in a separate ticket, if no such ticket already exists.

      Attachments

        Activity

          People

            charlie.swanson@mongodb.com Charlie Swanson
            charlie.swanson@mongodb.com Charlie Swanson
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved: