Registering server taking 3 minutes after a long pause from publishing

[BrianToussaint]

We are on Windchill 10.2 m030 CPS09.  This last weekend we switched out our publishing machine.  If there hasn't been a publish job in 15+ minutes and one runs, it is taking 3 minutes to register the server.  There is nothing in the logs that I can find.  It also happened with the first job of the morning when the workers were off.  We did go from one worker to two workers.  The first worker is the one that always catches at the registering server.  I can submit another job and worker 2 will register right away and work quickly.  Also if I submit another job within the allotted time, it will register right away.  I have turned off the "sleep" functionality of the network card on the worker.  The one thing that I do not know if it is a factor somewhere is that worker 1 and the old worker were configured to the same port from the Worker Agent.  Once Worker 1 has been connected, it is perfectly fine until it hits a certain amount of down time.  I am not shutting down the workers after an idle period.  Any insight on this would be much appreciated.

[BineshKumar1]

Are you using the same machine for both the workers? Do you have different cache locations defined for workers?

[BrianToussaint]

Yes it is the same machine for both workers.  Yes I have a different cache location for each worker.  I did notice this also when I was testing this server with just my test Windchill environment and thought it was because my Windchill test environment is really old, like it takes over 3 minutes for a method server to start.

Yesterday it happened in the middle of the workers being active.