Scaling Reggie
Dennis Reedy <[email protected]>
| Newsgroups | gmane.comp.java.sun.jini |
|---|---|
| Message-ID | <[email protected]> |
Hi, I've been asked to evaluate a deployment [*] that at the moment has 800 clients performing simultaneous discovery of a lookup service, with subsequent discovery of services that have active service registrations [**]. We are getting out of memory exceptions when the clients start up and do the discovery (this occurs within a 5-6 second window). Each client correlates to a physical machine. The immediate thought was to add a random wait for the clients to initialize, this may not be possible due to other (assumed) application constraints. Is the right place to look would be the com.sun.jini.reggie.taskManager setting? What would be a recommended setting for this property? Another approach is to partition the clients to separate lookup instances, perhaps 200-300 per lookup. Are there any rough ideas on what the "right" number here could be Has anyone scaled reggie to this level (or beyond)? Thanks Dennis [*] Sorry for incomplete info at this time, I dont have much to go on other then general description of the problem [**] There are approximately 300 services registered. -------------------------------------------------------------------------- Getting Started: http://www.jini.org/wiki/Category:Getting_Started Community Web Site: http://jini.org jini-users Archive: http://archives.java.sun.com/archives/jini-users.html Unsubscribing: email "signoff JINI-USERS" to [email protected]