I suggest running the RMI server in a separate JVM from the JCR server. When I implemented model 3 I had custom node types, so I had the server check if they existed when it starts up. It then creates the node types if they did not exist. Otherwise the server just made the repository visible over the network.
Hope that helps. -----Original Message----- From: hsp [mailto:[EMAIL PROTECTED] Sent: Tuesday, April 04, 2006 9:05 PM To: [email protected] Subject: About jcr-rmi in deployment model3. Have anyone experienced to build an application using the jcr-rmi for jackrabbit deployment model 3? Is there any issue, limitation, or something to avoid using such model? To do this kind of model possible to run, both side application must be implemented (client and server)? Or, all the implementation must be in the client side and in the server side will be only the configuration to make up the repository visible, provided by jackrabbit? Or would be the server side implemented with all needs and the client side only requests the services? Well, to sumarize, I would to know how must the application be made (refering to client and server sides in deployment model 3). Now I am querying you a lot, but consider that I want to contribute back after get better knows about. Thanks Helio
