We are migrating to a new box and trying to obtain sequential dumps of data.

Last Friday night, I created a job to generate sequential dumps and the job aborted Saturday at 3:12 AM. log.bshell reports that connection to server was lost. I am certain that the job was running before loosing connection. Also, there is nothing to indicate a power problem.

Is there a way to determine why connection to server was lost?

So nothing in your logfiles that indicate this?
What is your setting for session_timeout and/or ds_timeout_detect, ds_max_wait.

Is there some network error possibly involved? Did you run the session on the server or on a client PC? Was the PC locked (if it was the client you were running on), do you use some kind of APM (Advanced Power Management) to close down components during idle time and do these components come up correct and quickly when waked up.

Sometimes a right question gives the answer...

The session was running on client PC, not server. If there was a network error, where do I look to make this determination? The only message that I can find is in the log.bshell and on my desktop. I have posted a picture of my desktop in attachment.

My system administrator is out of town. How do I check the settings for session_timeout, ds_timeout_detect, and ds_max_wait?

I don't know if we use APM. I doubt it.

I do know that I have scheduled long job in the past. Typically archive sessions.