
The ansible service module is smart and will check the status of a daemon and only start it if deemed necessary when asked to start a service. The problem with this is that jenkins can go sideways confusing the daemon system making it think the service is up even though we have kill -9'd it into oblivion. The simple work around for this is to use restart which will always bounce the service in ansible. Change-Id: I240c318e9fa824d88bb817cf382d50ba2902d668
13 lines
562 B
YAML
13 lines
562 B
YAML
---
|
|
- hosts: 'jenkins0*.openstack.org'
|
|
# Treat any errors as fatal so that we don't stop all the jenkins
|
|
# masters.
|
|
any_errors_fatal: true
|
|
tasks:
|
|
- shell: '/usr/local/jenkins/bin/safe_jenkins_shutdown --url https://{{ ansible_fqdn }}/ --user {{ user }} --password {{ password }}'
|
|
- service: name=jenkins state=stopped
|
|
# This is necessary because stopping Jenkins is not reliable.
|
|
# We allow return code 1 which means no processes found.
|
|
- shell: 'pkill -9 -U jenkins || [ $? -eq "1" ]'
|
|
- service: name=jenkins state=restarted
|