slave does not upgrade cleanly

[esalberg]

When upgrading the jenkins slave jar from 2.0 to 3.8, the result was two running slave jar process (the old one and the new one).

I'm guessing that this is because once /etc/sysconfig/jenkins-slave is updated with the name of the new jar file, systemd isn't able to locate / shutdown the old process.

I could put together a PR that adds logic for upgrade where the service is stopped before /etc/sysconfig/jenkins-slave is updated (i.e. an if statement plus an exec), but I wasn't sure if there was a better way to do it.

RHEL7.4, Jenkins module commit 73bb6c9

[rnelson0]

I am unfamiliar with the best way to do this, but I have a request on the terminology. Per JENKINS-27268, terminology has changed from master/slave to master/agent. As the issue notes, this is complete in Jenkins 2.0 internals, but changes in plugins is still ongoing. I propose that we do our best to change the language from slave to agent where possible (it looks like some filenames may still persist, for example) while making changes to related code. Thank you.

[jhoblitt]

@rnelson0 We should probably open another ticket to rename the terminology in this module. However, note that usage of slave, node, and/or computer are still present in class names. api end points, and "scripted" pipeline.