so I was playing around with capistrano and a local dummy raring vm in vbox. everything was fine and my attention shifted to deployments on real machines. A couple of days later I want to deploy to my vm again and the script succeeds but afterwards no tomcat is running.
From all the messages that capistrano plays back to me, it looks like everything, including tomcat startup, went smoothly. Even the PID file was wirtten, but when you look at ps aux, the process doesn't exist. There is nothing in catalina.out but it's timestamp is updated, which is the only evidence, together with the PID file, that something happened on the vm at all.
Next thing I did was take the whole tomcat startup command as logged by capistrano and execute it "manually" with ssh. That gives me the exact same log messages but this time, the tomcat instance is actually up and running.
So, something that capistrano is doing with that command before it gets sent over via ssh must mess it up, but understanding the capistrano source of the run command is giving me a hard time, so any pointers would be appreciated.
