Improve how Jira handles requests to NFS Shared Home during NFS temporary outages

XMLWordPrintable

    • Type: Suggestion
    • Resolution: Unresolved
    • None
    • Component/s: Issue - Attachments
    • None
    • 0
    • 2

      Issue Summary

      During a temporary NFS outage, the Tomcat threads hang indefinitely and gets exhausted.

      Steps to Reproduce

      1. Configure NFS Shared Home
      2. Install Jira Data Center
      3. Breaks NFS communication

      Expected Results

      • Jira should detect NFS communication failure.
      • Requests relying on NFS should fail faster.
      • Requests NOT relying on NFS should continue working fine.
      • Threads should not hang indefinitely.

      Actual Results

      • Tomcat threads hang indefinitely
      • Tomcat thread pool gets exhausted
      • All nodes are affected, causing Jira downtime. 

      Additional Information

      With the NAS down, the operating system keeps the volume mounted. 
      Any OS commands like LS, CAT would also hang. 

      Neither JVM nor Tomcat has a built-in mechanism to directly limit the total processing time of a request.

      Workaround

      Currently there is no known workaround for this behavior. A workaround will be added here when available

            Assignee:
            Unassigned
            Reporter:
            Bruno
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

              Created:
              Updated: