Description of problem:Customer has repeatedly encountered the httpd service being stopped for no identifiable reason. Thsi causes the UI workers to fail and leaves UI worker appliances essentially zombied. Version-Release number of selected component (if applicable): 5.6.3.3 How reproducible: cause is unknown and is the reason for thsz BZ. symptom should be recreate-able by terminating Apache server on UI appliance Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
google drive is being prepared to provide most current logs provided by the customer.
link below is the corrected shareable link to the folder "BZ 1448766 - SF 01724882" which contains all of the customer provided diagnostic materials https://drive.google.com/drive/folders/0Bw3u5g60tmjud2FSbElwbDhOaTQ?usp=sharing
PR: https://github.com/ManageIQ/manageiq/pull/15078
New commit detected on ManageIQ/manageiq/master: https://github.com/ManageIQ/manageiq/commit/5a8f28e32105e54c1b251142c4c105fb89bec7da commit 5a8f28e32105e54c1b251142c4c105fb89bec7da Author: Gregg Tanzillo <gtanzill> AuthorDate: Thu May 11 19:44:11 2017 -0400 Commit: Gregg Tanzillo <gtanzill> CommitDate: Thu May 11 19:44:11 2017 -0400 Start Apache if roles were changed and it is needed by the current roles Fixes a bug that was introduced in da9523ee2168da89511c9260528c7bb243bfb777 when the Apache load balancer was made static. That removed the code that started Apache up after the role or worker count changed. https://bugzilla.redhat.com/show_bug.cgi?id=1449766 app/models/miq_server/environment_management.rb | 4 ++++ app/models/miq_server/worker_management/monitor.rb | 2 ++ 2 files changed, 6 insertions(+)
Verified on 5.9.0.1.