Well, at least on a new setup I built it is, and with the same symptoms as JENKINS-7641 <https://issues.jenkins-ci.org/browse/JENKINS-7641>.
It is only happening on a certain cluster of nodes, all of them cloned from the same original VM template: - Everything is on CentOS 6 and hosted in VmWare. - The jobs where this is occurring are all copying large artifacts such as ISO images. - Jenkins home area is NFS-mounted - The problem does not surface on every build. I suspect it builds up after the node is disconnected/re-launched. - Stuck builds are not even killed by using the Build Timeout plugin. The only way to abort the job is to disconnect the slave and then abort the job via the UI. So far I have ensured that everything uses Oracle JDK7, I have reduced the executors to 1 per node, and have set -Dhudson.remoting.Channel.pipeWindowSize=8192 on the slaves. I have the problem reproduced on two independent installations, but utilising the same kind of nodes, so it cannot just be one bad master. As of now I have one build node that has a stuck build on it and has been in that state for 13 days. I have left it so in the hope that I can figure out what debugging information to retrieve from it. Any pointers to where I can start debugging this? THe normal stack dumps on the node System Information page looks pretty normal. -- 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/groups/opt_out.
