We have a master node running on Linux (AWS), which we use to start up Windows nodes (also AWS). We use the Slave Setup Plugin (https://wiki.jenkins-ci.org/display/JENKINS/Slave+Setup+Plugin) to start nodes on demand; however, jobs are not being scheduled because Jenkins sees all nodes offline. As I understand it, this plugin only starts nodes if there are jobs scheduled. Clearly, there is an inconsistency between Jenkins and this plugin.

What I expect to see is that jobs get schedule, which causes the slaves to be started on-demand and then they get built.

What I actually see is that Git Polling reports the following:

Git Polling Log

Started on Aug 19, 2016 10:46:00 AM

We need to schedule a new build to get a workspace, but deferring 68,624ms in the hope that one will become available soon (all_suitable_nodes_are_offline)

Done. Took 0 ms

No changes

I can confirm that these nodes have workspaces and if I trigger a job manually the process works as expected.

This is a critical issue for us as we cannot have AWS instances running idle as they are far too expensive. We only want them to be started when there is something to build. We have configured the nodes according to the instructions of the plugin (please see screen shot)

Version of Jenkins: 2.18
Version of plugin: 1.10

I've attached full details as PDF files.