We have ran into quite a few bugs attempting to upgrade the EC2 Plugin from 1.39 to 1.42.
I want to report a bug about them, but i want to understand more before i do, so i can raise those bugs correctly.
I can see that the EC2 Plugin is handling the provisioning of windows slaves through winRM.
WinRM streams the slave.jar file to the temp directory and then issues a command to start that.
But when it comes to issued bat commands from the Jenkins pipeline itself i believe that the communication there is purely between the master and the slave through the jenkins remoting layer (slave.jar to Jenkins master) and winRM has no involvement in the issuing of bat commands. is that an accurate statement?
Our issues are faced typically half way through the execution of jenkins pipeline when we get timeout exceptions. I want to know if this could have something to do with winRM so i am asking this question first to see if WinRM is at all involved in the pipeline bat executions.
Also, since the java slave was kicked off by winRShost, what would happen if winRS process died. would this kill the java process as it is a child of winRS?