New operation abort and restart

[stefanbrausch]

This pull request introduce the new operation "abort and restart".

Motiviation for the change: Some of our testing jobs have problems with timeouts. In this case the job would be aborted and restarted now.

[oleg-nenashev]

For me it smells like a duplicate of Naginator's plugin functionality. I'm OK with it, but probably it makes sense to integrate plugins somehow.

[ikedam]

You'd be interested in how [Rebuild plugin|https://wiki.jenkins-ci.org/display/JENKINS/Rebuild+Plugin] reruns builds.

• That preserves build parameters.
• That preserves build causes. * This preserves relation between upstream builds and allows retriggered builds to copy artifacts again. * Additionally, rebuild plugin displays very useful link "This build is a rebuild of #xxxx".

[oleg-nenashev]

@ikedam I'd vote for Naginator, but the reason is same. Rebuild seems to be more healthy, so I will agree with such approach as well

[ikedam]

I agree that "fail the build" and [Naginator plugin|https://wiki.jenkins-ci.org/display/JENKINS/Naginator+Plugin] provides the same functionality. Though that doesn't make this request rejected immediately, it's prefered that this feature is useful than Naginator plugin in some points.

[oleg-nenashev]

@ikedam I agree, such strategy would be useful. I'm mostly aware about the backend mplementation

[stefanbrausch]

First: The reason why we are use build-timeout and not periodic reincarnation (it's similar to nagiator plugin) are the email notification. We've more then 25 (Fitnesse) Jobs per night with this problems, and a lot of failing emails. We prefer an aborted build instead a failed build also. Second: thanks for your comments. @Jochen-A-Fuerbacher will create a new pull request.

[stefanbrausch]

@Jochen-A-Fuerbacher has created a new pull request

[ikedam]

JFYI. You don't need to create new pull requests for new changes. You can update a pull request by pushing new commits to your branch.