Automatically re-running a remote job without failing it, if the connection to the node drops
I see jobs failing with "FATAL: command execution failed" caused by java.io.EOFException. I understand the connection to the remote node drops, so java.io.EOFException is raised. Is it possible to re-run the job remotely without failing it at all? I've seen the Naginator plugin, but it seems to start a new separate job if the current one fails, which is not what I want. The failure would bubble up to the parent job which started it and cause it to fail.
Or is it possible for the parent job to check the status of the child jobs started by it, and "manually" restart the ones failing due to java.io.EOFException until they succeed?
Building remotely on instance-1 (tag1) in workspace /var/lib/jenkins/workspace/eval