for 1:There are several process. Especially server and user are proofing each other there certification, to accept or denie the connection. What is the background for your question?
for 2:
You can’t trigger client-server-replication by an event like ‘when mail comes in’. You have to choose a short replication interval (may be 10min) in the used locationdocument, to get mail quickly in your local replica.
for 3:
If you have a file-backup of your Admin-ID with the initialized password, you can use this ID-File (if you know the initialized password). In other cases, you need a second ‘Admin’. If you have activate the recovering process, he can extract your recovering password and you are able to set a new password for your ID. This is the clean solution. Last possibility is, the second ‘Admin’ will register a new ID for your admin account. This is not commandable, but sometimes if the system is implemented quick and dirty, there is no alternative. If all this did not help, you have a real problem.
Lots of things happen, the user is authenticated by the server, certificates checked for expiry, the servers calculates all the access groups the user has and caches them on the client, replication is initiated if it is set to run, cluster parameters are gathered if the user is in a clustered environment. There are more, but that is a start.
If you want this, why not have the user work off the server, but have a local replica to use when disconnected… You can set replication to occur every minutre if you wanted, but that would create overhead on the network for little benefit.
You don not need an admin id or password to use the Admin client. The id that wants to use the Admin client needs to be in the “administration” fields in the server document.