Activity Stream Gadget or Project Activity loads blank after a Jira restart.

XMLWordPrintable

    • Type: Bug
    • Resolution: Unresolved
    • Priority: Low
    • None
    • Affects Version/s: 9.12.22, 10.3.3, 10.3.5, 10.5.1
    • Component/s: Dashboard & Gadgets
    • 9.12
    • 4
    • Severity 2 - Major
    • 3

      Issue Summary

      Activity Stream Gadget or Project Activity loads blank after a Jira restart.

      Steps to Reproduce

      1. Install a vanilla instance of Jira Software Data Center.
        This was validated on JSW DC versions 10.3.3, 10.3.5 and 10.5.1.
      2. Configure a second Tomcat connector to bypass the Load Balancer
      3. Set the legacy mode to enable logging directly through node. Insert the variable -Datlassian.authentication.legacy.mode=true in the <jira_dir_install>/bin/setenv.sh file, inside the JVM_SUPPORT_RECOMMENDED_ARGS parameter.
      4. Start Jira
      5. Access Jira through second connector, using http://<nodeIP>:<port>
      6. Restart Jira application while the browser tab is still open using the second connector

      Expected Results

      Activity Stream gadget should use the browser address as reference to send requests to load data.

      Actual Results

      Activity Stream gadget will cache the second Tomcat connector address and use it to send requests to load data.
      It happens despite the address URL being used in browser address bar. Even using Jira Base URL.

      The problem may affect nodes individually.
      While some nodes may be affected, others may work normally.
      Restarting the nodes may fix the issue, or cause it if someone access the node using the alternative connector.

      The below behavior can be observed in HAR file while browsing any screen containing an Activity Stream gadget:

      The below behavior can be observed in gadget while browsing any screen containing an Activity Stream gadget:

      Activity Stream Gadget in a Dashboard:

      Activity Stream in a Project Summary:

      Workaround

      The affected node can be restarted individually.
      Avoid accessing nodes directly using <nodeIP>:<port> after a restart.
      Wait for a few minutes (from 2 to 3) after it comes online to access the alternative connector and bypass the load balancer.

            Assignee:
            Unassigned
            Reporter:
            Bruno
            Votes:
            3 Vote for this issue
            Watchers:
            8 Start watching this issue

              Created:
              Updated: