Hi all,
My company has a server that has uhm... served us well for years, but we 
are now developing an Apache/Tomcat/PHP/MySQL application that bogs it 
down somewhat.

(Netfinity 5000, dual PIII 550MHz, 832MB RAM, Ultra160 SCSI disks)

Right now, when we use the administrative pages, one CPU gets very busy, 
so it's easy to see it's time for an upgrade.

However, in the future, it may not be the CPU but say, disk I/O or memory
bandwidth causing the slowdown. And I don't even know how to find out
which might be the cause. Are those details logged? Can I use some log
tracker or monitoring system to say things like this:

Hey! The CPU was 100% busy for a stretch of 300 seconds, and spent 1.2% of 
the day at full use.

Hey! There were a maximum of 12 processes waiting for disk today, and a 
stretch of 30 seconds with more than 1.

Or can I use something like top to view disk waits, memory utilization and 
CPU usage? Oh and network bandwidth? Our server is behind a T1 and that 
may become a bottleneck too.

Chris


_______________________________________________
TCLUG Mailing List - Minneapolis/St. Paul, Minnesota
http://www.mn-linux.org tclug-list at mn-linux.org
https://mailman.real-time.com/mailman/listinfo/tclug-list