Yaml-ized jobs: Can they be unstable?

David Caro Estevez dcaroest at redhat.com
Mon Apr 25 19:38:47 UTC 2016


With the standard ci is not possible, as the idea of 'unstable' and 'failure' is quite ambiguous and usually leads to confusion.

So we decided not to have that third state, runs either pass, or do not.

You can print/archive/log anything you want to allow you help debugging the issue, but on the ci side, the decision is to -1 or not a patch, so just two states.

Out of the standard you can define post-build scripts that can modify the state of the job (you can set it as failed, unstable or even pass a job that otherwise would be marked as failed).

Though for the reasons I exposed, I don't recommend that.

David Caro

El 25/4/2016 8:30 p. m., Fabian Deutsch <fdeutsch at redhat.com> escribió:

Hey,

is there a way how a yamlized job can become unstable?

I'd liek to run some sanity node tests after a run, and mark a run as
unstable if this happens.
According to the jenkins docs, a job is unstable if one or more publishers fail.

Can this be achieved with yamlized jobs?

- fabian
_______________________________________________
Infra mailing list
Infra at ovirt.org
http://lists.ovirt.org/mailman/listinfo/infra
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ovirt.org/pipermail/infra/attachments/20160425/35386da2/attachment.html>


More information about the Infra mailing list