-
Improvement
-
Resolution: Unresolved
-
Major
-
None
We regsietered all out vm in jenkins an want to do thing on thme by job. At the moment the vm is tried to power on periodeially (dunno what this depends on) and we want only that the machine will be tried to start if the job needs it.
So not only "do nothing" if its disconneted or don't do anything if its connected furthermore only try to switch it on (power on) and connect it, if any job requires it (by slave name or tah label). I already set option "use this node exluively for assigned jobs" and "on/off on demand/idle" but still the node is automatically swicthed on, although no job runs request it. Its impossible to power down the slave doing maitenance, although the jenkins admin do make it temporatly not availabe before. That is more a woor around than the wanted schedule typ i request in this issue.
[JENKINS-20011] new slave schedule typ: only power on machine if job requests it
Jason Swager
made changes -
| Assignee | Original: Stephen Connolly [ stephenconnolly ] | New: Jason Swager [ jswager1 ] |
As i said i switched on "use this node exclusively for assigned jobs" and no job is assigned. Btw it would be helpful to see "exclusively" and "possible assigned jobs (by tags)" in node view.
Jason Swager
added a comment - Sorry - the vSphere plugin uses the standard Jenkins controls/logic for how jobs are tied to a slave. By chance, did you look at the history of the slave to see if any jobs had been assigned to it. Or was there anything in the log?
Finally, did the slave get started at the same time as another job was run? There is logic in the vSphere plugin that will start the slave if Jenkins queries that slave with a queued job and the job CAN run on the slave. In the first versions of the plugin, this would cause EVERY VM slave that matched the job to startup. That was later modified so that only one slave at a time could startup.
Jason Swager
added a comment - Sorry - the vSphere plugin uses the standard Jenkins controls/logic for how jobs are tied to a slave. By chance, did you look at the history of the slave to see if any jobs had been assigned to it. Or was there anything in the log?
Finally, did the slave get started at the same time as another job was run? There is logic in the vSphere plugin that will start the slave if Jenkins queries that slave with a queued job and the job CAN run on the slave. In the first versions of the plugin, this would cause EVERY VM slave that matched the job to startup. That was later modified so that only one slave at a time could startup. Jason Swager
added a comment - Any further information in this issue? If nothing new, it will be closed.
Jason Swager
added a comment - Any further information in this issue? If nothing new, it will be closed. Jason Swager
added a comment - No additional information provided. Closing as non-reproducible. Jason Swager
made changes -
| Resolution | New: Incomplete [ 4 ] | |
| Status | Original: Open [ 1 ] | New: Resolved [ 5 ] |
i am using 1.1.0 and no job is tied to this node, but as soon as i shutwown jenkins powers on the node again immidiately and no jobs start afterwrds on this node, as there is none defined to do so.
| Resolution | Original: Incomplete [ 4 ] | |
| Status | Original: Resolved [ 5 ] | New: Reopened [ 4 ] |
Jenkins will send a job to the slave only if necessary. It will prefer one that is already powered on over one that is powered off. This feature is already in heavy use without issues. Are you sure that there isn't any job - even a flyweight one - that is trying to run on the slave that would cause it to power on?