I think traditionally this kind of use case has been implemented as a Cloud extension point and RetentionStrategy (or perhaps QueueTaskDispatcher.)

The idea is to create a new slave and then throw it away immediately after use.

Jenkins maintains a connection to a slave all the time, so rolling back a slave after a job is assigned wouldn't be very practical.

But if you find a way to make it work, good!

On 08/07/2013 01:21 PM, Alex Szczuczko wrote:
Hi,

I'm trying to improve libvirt-slave-plugin so that it can revert a snapshot 
before a job runs on a node. I've tried using several extension points to do 
this; the best I can do (With RunListener) is revert the snapshot and have the 
job fail because the node connection is lost (reconnecting doesn't seem to save 
the job).

As far as I can tell, what I need is access to the point where a job is 
assigned to a node, but not yet dependent on that node's connection. I have 
been unable to find this point myself, does anyone know of anything like that?

Thanks,
Alex



--
Kohsuke Kawaguchi | CloudBees, Inc. | http://cloudbees.com/
Try Jenkins Enterprise, our professional version of Jenkins

--
You received this message because you are subscribed to the Google Groups "Jenkins 
Developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.


Reply via email to