service state=restarted: Documentation wrong? Or behaviour wrong?

Archives Ansible Project
1

“Service” documentation indicates “restarted” is not idempotent:

‘started’, ‘stopped’, ‘reloaded’, or ‘restarted’. Started/stopped are idempotent actions that will not run commands unless neccessary. ‘restarted’ will always bounce the service, ‘reloaded’ will always reload.

Playbook:

action: service name=hadoop-hdfs-namenode state=restarted

Actual behaviour:

failed: [10.20.30.40] => {“changed”: false, “failed”: true, “msg”: “failed determining the current service state => state stays unchanged”}

Aside: this is during installation/configuration of HBase, which seems to have broken “service” scripts and will possibly never return proper state. Would like to be able to install HBase anyway.

2

sounds like you were using upstart and need to update

https://github.com/ansible/ansible/issues/1067

3

Thanks for the URL. I’ve read it with great fascination and I think you’re right. I’m on the latest Ansible now (but I also think I worked around the problem using a pattern=).