I'm attempting to use Scripted Cloud plugin to orchestrate system tests for a product. For most part I was able to achieve my goal. Only remaining problem is that Scripted Cloud plugin fails to shutdown and restore virtual machine in time if another test (job) is already pending in build queue. In such case Jenkins slave on VM managed by Scripted Cloud plugin reconnects to master before plugin is able to shutdown VM. Queued test starts but is interrupted by virtual machine shutdown only seconds after it starts. Slave is configured to use Java Web Start. Attached log should illustrate the problem.
I was unable to find solution for this problem. Is there any way to prevent queued job to start execution before Scripted Cloud plugin is able to shutdown VM? My guess is that Scripted Cloud plugin should somehow instruct slave to stay disconnected and reconnect only after it is restarted by container. Marking slave as temporarily offline is probably not a good solution, since this feature is needed to keep slave offline for VM maintenance. Acceptable workaround would be to increase reconnect delay for a slave, but I cannot find any documentation about that. Current reconnect delay seems to be 10 seconds. Connection from slave is reset only 16 seconds after reconnect. Increasing slave reconnect delay to 60 seconds would probably solve problem for me. -- 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-users+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
jenkins.err.log
Description: Binary data
