I think you don't have any problem, that is normal.
You can use Resume command/option to change state to running.
If you want, check [http://edocs.bea.com/wls/docs92/server_start/server_life.html#wp1052817]
Is it normal??? Then one of us does not understand something. Then once again. The server is starting into RUNNING state. Then it is ready to handle operational requests. After some time it goes to ADMIN state without any interaction from the user's side. How can it be normal??? In such a case the server can't handle operational requests and has to be resumed manually. Of course i can do that in a lab or test environment, but what if such a situation happened in a production environment?
Well, I've had the same situation when I'm working in development mode , I think managed servers are not working and themselves changes its state.
In production mode I have not had that situation. Probably I've had good look.
Well... If it works well in production mode, then I'm OK. But IMHO, even in development mode, such a behaviour is far from normal. How to develop and test effectively if the server goes off unexpectedly? I think I will report a case to the support.
i encountered similar issue while trying to start my managed server which hosted BI publisher on bi cluster.
reason: the datasource pointed to this cluster was not able to connect to the database(the db was down).
solution: bring up the database and restart the server.