Planning For Disk I/o

by Jun 13, 2007

I am just starting to get into capacity planning. Does anybody have any words of wisdom around planning for disk I/O growth based on deployment of a new application to a production server? I am working on an OS where I can't view disk I/O by process, that would let me run the new app on a test box and measure the I/O that way before rolling it out. The system is running about 40% busy with spike up to 80% during peaks, I am worried that the combined peak of the new application will kill my users sessions.

Any help is appreciated.