[jboss-user] [JBoss Cache: Core Edition] - Re: JBoss cache instances fail to join cluster after bounce

setatum do-not-reply at jboss.com
Wed May 6 13:02:20 EDT 2009


Well I just realized something that may be important. Originally I had tried using Pojo Cache for our implementation but ran into serious memory leak issues. I switched the implementation to use core cache, and this got rid of our memory leak problems.

However, the config file we use now is based on the config file from section 8.2 of the Pojo Cache User's Guide. Looking it this and comparing it to the configuration files from section 12.1 of the Core Cache User's Guide, it is totally different! However, core cache seems to load this configuration file ok (if I intentionally mess up the file format, the loader complains).

I'll now attempt to setup a new configuration file based on the core cache examples.

Is there a tuning guide separate from the User's Guide which recommends settings based on your setup? For example, my setup is 3 servers all close together on a high speed network, where speed is more important than throughput (our cached datasets are currently very small). The instances themselves should be reliable, but if an instance properly shuts itself down with stop/destroy for ~5 minutes and then attempts to rejoin the group, it should be allowed to do so (this seems to be my problem now).

-Scott

View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4229116#4229116

Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4229116



More information about the jboss-user mailing list