Right now, we are using jcs for some internal caches, of which we are replicating 3 caches to an oracle database for jvm failover. We have this set up where we write to the in memory cache, and have an event queue thread pool configured that updates the database asynchronously. This works very nice for the most part. The problem is we have many tasks per application page that update the cache, so we end up running a lot more updates to the database than really needed for the failover purpose. So an example- 3 pages in the application, we calculated 60 puts to the cache from jcs admin page. So I am interpretting that as 60 sql statements for 3 pages. In our production environment, that will translate to 1000's of sql statements per second and add undesired and unnneeded load on the database servers. I was wondering if there was some way to configure the jdbc cache to just be updated every 10 or 20 seconds with one update from memory that would contain all intermitten updates for that time period. In other words- we do not need the database as up to date as the in memory cache. If you are familiar with http session configuration (in WebSphere)- the concept is very similar. Is there a way to configure jcs with this in mind, or would we need to modify the jcs code base? (By the way- the application code is vendor application code and design changes there are not really an option). -- View this message in context: http://www.nabble.com/pooling-jdbc-updates-tp20154451p20154451.html Sent from the JCS - Users mailing list archive at Nabble.com.
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
