Ok - so the problem with the immediate death of the process was that I was not using the correct shell. I was using the default /usr/bin/sh from a remote terminal. When I exited my shell, the process died. I now use the bash shell and this does not happen.
However, that was the latest problem - the other problem whereby the server dies unexpectantly and quietly after some unspecified period is still outstanding. As of right now, the server has been running for 24 hours with no problems. I have two instances - one has my WAR file deployed, and the other does not and both continue to run. Although I have not tried using them in earnest yet, obviously the one without my WAR file, I cannot use other than the admin console.
Additionally, we realized that there *may* have been a conflict with one of the ports being used for the default JMS host. It defaults to 7676. We didn't see any other domains using this port, but a colleague said they had seen similar behavior - aka a domain dying quietly if another instance/process comes along and wants to use that port. Now - as I said, we don't know of anyone needing that port number AND I do not even use the default JMS provider - it is just the default that is part of the Glassfish instance.
I also found out that the server is overloaded and maxing out on resources but then I thought I would have seen OOM errors or something similar.
I am about to update my JDK to the latest version to see if it removes my dependency_failed messages.
That is where we stand now. I will keep you posted if/when I have more information.
Thank you
Brenda
[Message sent by forum member 'bcoulson220']
http://forums.java.net/jive/thread.jspa?messageID=393342