> From: Juha Laiho [mailto:[EMAIL PROTECTED] > Subject: Re: Xmx and Xms size > > This recommendation comes from seeing (albeit with a now > obsolete JVM) excessively long pauses for GC in a situation > where an application with memory leak had gradually been > given higher and higher Xmx values
That's not a terribly useful generalization, given that the webapp you're describing could never reach anything approximating a steady state due to the memory leak. Current JVM GC pause times are a function of the number of live objects, not the size of the heap. Since the number of live objects tends to be a constant for a given load, the pause time is also constant. The largest heap possible is usually the right answer these days, at least until it impacts on other applications running on the same system, or you start encountering non-uniform memory access (NUMA) situations. The latter can significantly impact performance on larger multiprocessor systems. - Chuck THIS COMMUNICATION MAY CONTAIN CONFIDENTIAL AND/OR OTHERWISE PROPRIETARY MATERIAL and is thus for use only by the intended recipient. If you received this in error, please contact the sender and delete the e-mail and its attachments from all computers. --------------------------------------------------------------------- To start a new topic, e-mail: users@tomcat.apache.org To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]