Ken, I think I have a bit of an understanding about virtual arrays. Unfortunately allocation doesn't, and shops without PAV still suffer from IOSQ.
And I still open the doors up to see if there some sort of IO skew happening - flashing lights still tell you something about destage and cache miss activity. Ron > -----Original Message----- > From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf Of > Klein, Kenneth > Sent: Tuesday, June 23, 2009 1:00 PM > To: [email protected] > Subject: Re: [IBM-MAIN] SMS Dataset Allocation Problem > > Most people are running on VIRTUAL disks anyway. z/os may "see" a 2105 > or a 2107 but in most cases its tiny little brown round and spinning in > RAID 5 on the SAN so it doesn't much matter any more. We used to place > datasets on disks in accordance to their activity level and distance > from the VTOC. To check your work you could open the door, look into the > HDA and watch how much the actuator was bouncing around. > > > Ken Klein > Sr. Systems Programmer > Kentucky Farm Bureau Insurance - Louisville > [email protected] > 502-495-5000 x7011 > > -----Original Message----- > From: IBM Mainframe Discussion List [mailto:[email protected]] On > Behalf Of Ron Hawkins > Sent: Tuesday, June 23, 2009 3:52 PM > To: [email protected] > Subject: Re: SMS Dataset Allocation Problem > > Ted, > > It was not DB2 that had a problem. It was CICS and a few batch jobs. I > don't have a problem with multi-volume datasets at all, and in fact I > design SMS to allocate large datasets on as many volumes as possible. It > was simply having so many candidate volumes for every dataset, 1 track > or 1 pack, was excessive and we started having TIOT blow outs. I think > we were using 20 for the unit count when the problem happened. > > We ended up using 5 for both VSAM and PS and, let SRS from DTS look > after adding volumes beyond five. > > Ron > > > -----Original Message----- > > From: IBM Mainframe Discussion List [mailto:[email protected]] On > Behalf Of > > Ted MacNEIL > > Sent: Tuesday, June 23, 2009 12:21 PM > > To: [email protected] > > Subject: Re: [IBM-MAIN] SMS Dataset Allocation Problem > > > > >TIOT > > > > Set it to the max, and (in general) it won't be an issue. > > And, IIRC, DB2 uses a different type of table for allocation. > > The number of datasets it can allocate is way beyond what the TIOT, in > theory, > > allows for. > > > > (This was about DB2, originally). > > - > > Too busy driving to stop for gas! > > > > ---------------------------------------------------------------------- > > For IBM-MAIN subscribe / signoff / archive access instructions, send > > email to [email protected] with the message: GET IBM-MAIN INFO > > Search the archives at http://bama.ua.edu/archives/ibm-main.html > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, send > email to [email protected] with the message: GET IBM-MAIN INFO Search > the archives at http://bama.ua.edu/archives/ibm-main.html > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to [email protected] with the message: GET IBM-MAIN INFO > Search the archives at http://bama.ua.edu/archives/ibm-main.html ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
