Uploaded image for project: 'Red Hat OpenStack Services on OpenShift'
  1. Red Hat OpenStack Services on OpenShift
  2. OSPRH-17944

[fork vs multiprocessing] 6 - Harden logging configuration against fork-related deadlocks

XMLWordPrintable

    • Icon: Story Story
    • Resolution: Unresolved
    • Icon: Critical Critical
    • None
    • None
    • python-oslo-service
    • None
    • Important

      Goal: Audit default logging setup: ensure any QueueHandler/QueueListener pair is re-initialized in child processes or guarded behind spawn context, eliminating the lock-held queue scenario from the blog.

      Acceptance Criteria:

      • Verify listener thread always starts after any fork (documented or via spawn ignoring).
      • Functional test: emit 1000 log lines during fork; child processes flush without hanging.
      • Update logging section of documentation to mention spawn requirement.

      Additional Details:
      See https://pythonspeed.com/articles/python-multiprocessing/ and video explanation at https://youtu.be/RIc-Tut95YM?t=185.

              mbollo@redhat.com Daniel Mats Niklas Bengtsson
              hberaud Hervé Beraud
              rhos-dfg-pidone
              Votes:
              0 Vote for this issue
              Watchers:
              1 Start watching this issue

                Created:
                Updated: