Finding problems on DASD volumes
Tuning a DASD system using tools like RMF is, to use a well-worn phrase, like drinking from a fire hose. These monitoring and reporting tools usually give their information by logical volume and since there can be thousands of volumes on a mainframe using DASD storage, the ones that need tuning can be lost in the flood of information in the reports.
The logical move is to use the sorting facilities of RMF or other tools to narrow the field to just those volumes that show numbers that are too high. The problem is that single numbers, even composite ones like the I/O Intensity (response time multiplied by I/O rate), don't necessarily pinpoint volumes in need of tuning. For example, says Tom Beretvas, of Beretvas Performance Consultants in Draper, Utah, a high I/O Intensity can arise strictly from a high transaction rate where the response time per transaction is well within acceptable limits.
Beretvas suggests a multi-step method of winnowing the RMF reports to focus on what he calls the "interesting DASD". The first step is to eliminate all volumes with response times below a threshold time, such as 7 ms, and low I/O rates, such as 1 I/O per second. The second step is to sort the remaining volumes by I/O intensity and concentrate on the remaining ones with the highest I/O intensity. Volumes with a combination of high I/O rates and slow response times that still show high I/O intensity are probably ones that can benefit from tuning.
Rick Cook has been writing about mass storage since the days when the term meant an 80K floppy disk. The computers he learned on used ferrite cores and magnetic drums. For the last twenty years he has been a freelance writer specializing in storage and other computer issues.