> From: David Delbecq [mailto:[EMAIL PROTECTED] > Since you say CPU usage on your machine is around 5% during > those tests, > and your database is on same machine as tomcat server, i doubt the > bottleneck really is on database.
The OP never reported on disk I/O - this is often the killer on database speed, and "80g HDD" to me suggests there's a single, fairly slow-spinning drive in the box. > The first ting i'll suggeste is to > take all queries run by your application, make an sql script of them, > and run this script using a mysql client. If the script > returns in less > than a second, it's your application the bottleneck at some > point. If it > takes 10 seconds, improve your database designs, add indexes, reduce > queries. Agree. Also check the disk I/O load - I didn't see which operating system the OP is running, but check Physical Disk / Average Disk Queue Length on Windows (in the Performance tool) or use "vmstat" on most UNIXes to get a rough-and-ready feel for the disk traffic. I'm not saying it *is* the problem, but I've seen far too many systems where disk ends up being the bottleneck! See "Amdahl's Rule of Thumb" near the bottom of http://en.wikipedia.org/wiki/Amdahl's_law - a balanced system needs 1 byte of RAM and 1 bit per second of I/O for each instruction/second the CPU can run. These days, the CPUs run 10^9 instructions per second, we have 10^9 bytes of RAM in a system, but most physical disks peak at around 10^8 bits/s of I/O, or under 10^7 if they're subject to heavy random access (say by a database). It's not enough! - Peter --------------------------------------------------------------------- To start a new topic, e-mail: users@tomcat.apache.org To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
