Polling without workspace failed to wake up offline slave node

[Jinny Pan]

Hi all,

I have a build job having a scheduled polling. It is configured to build on a slave node. The node's availability is set as "Take this slave on-line when in demand and off-line when idle". 

Ideally, whenever new code come in, the polling would trigger a new build, the new build will be put into a queue, then the slave node would wake up automatically to start build.

But the reality is, even there is new code, the polling seems doesn't work, the pooling log says: 

We need to schedule a new build to get a workspace, but deferring 862,449ms in the hope that one will become available soon (all_suitable_nodes_are_offline)

  Done. Took 0 ms
  No changes

The build queue is empty. The slave node never wake up.

Does anybody has the same issue before? I am using Jenkins ver. 1.625.2 

Thanks in advance!

[John Mellor]

I always have this issue whenever the master has been taken offline and restarted, and occasionally when recovering from a slave connectivity issue.  I have to walk through the slaves and manually bring them online once before they work correctly.

 

There are multiple bugs in the way that Jenkins keeps track of its slaves, recovers slave state from downtime, and polls slaves when it should not.  I think this is another one.

--
You received this message because you are subscribed to the Google Groups "Jenkins Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to jenkinsci-use...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/jenkinsci-users/9ff29e1b-363a-42de-89de-10c8d0579f92%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.