I'm working on such a tool for some of my customers: it analyses PTK's Extended Summary files and you can then extract whatever you want (CPU, page rate, IO rate). But, my tool is meant to make GDDM graphics from it (requires GDDM/REXX). Creating a CSV file is an option I want to provide too.
If you would have used my PRFGUI tool, you know what to expect. Anyhow, you could extract some code from the REXX code I use and save the created numbers in some other file. I've almost ready something I call version 0.8. You want to have a look? You can use CMS Pipelines? Another alternative: have PTK create a Trend record every 10 minutes (FC MONCOLL RESET), but that would result in huge FCXTREND files. But, PTK allows you to make graphs from it. 2010/7/23 Michel Beaulieu <[email protected]> > Good day, > > I am trying to do something that seems to me a good challenge to do > with z/VM performance toolkit. > > The end result that I want to have is a set of records for a selected > number of userids. > > Each record would be for a 10 minutes interval. > > format: > yyyymmdd hhmm userid cpusec > > That means I would create 24 x 6 = 144 records per day per virtual > machine. > > Accumulating that over many months, I can load that data in my favorite > spreadsheet > and do all kind of numerical analysis. > > I am trying to use z/VM Performance Toolkit to help me to do that and I am > stuck! > > The best I could do was to use plotvar with trend files. > However, that is too much aggregated for my need. > > I welcome your suggestions. > > Michel Beaulieu > Montreal, Canada > > -- Kris Buelens, IBM Belgium, VM customer support
