Mike, I didn't mention Cache partitioning :-) I don't think it is appropriate for the problem you have described so far.
I'm confused that you are saying the DB2 IO is low, when you previously described these as "fairly active disks." If the consolidation was a folding of whole volumes, then you need to recalculate the old cache hit rates from the consolidated counters. Averaging the cache hit rates of the old volumes is not valid unless the IO rates and types are similar. If the consolidation is folding datasets from many volumes I suggest you use the SMF Type 42 subtype 6 record to check the before and after cache hit rates. One random thought is to check your MSR settings in the STORCLAS if these are SMS managed. The 3390-27 is regarded as a 3390-9, and the default MSR used by DCME is different to a 3390-3. *** This is a wild arsed guess *** I'm hopelessly out of date as to how DCME works with blank MSR fields nowadays, but it's something I thought customers should be looking at when considering volume consolidation. Ron > -----Original Message----- > From: IBM Mainframe Discussion List [mailto:ibm-m...@bama.ua.edu] On Behalf Of > Mike Shorkend > Sent: Tuesday, December 29, 2009 3:37 AM > To: IBM-MAIN@bama.ua.edu > Subject: Re: [IBM-MAIN] HDS/USP Model 27 DASD and DB2 > > Thanks for your responses. > Ed/Adolph - the disks are net new - we have not started consolidating yet. > > Ron - Cache disk ratio was my error - I meant cache hit ratio. > > The DB2 disks used until we started this experiment are all MOD-3 with > hyperpav. > > I am not aware of cache partitioning - I will look into it. > > We currently have 3 MOD-27s and approx 190 MOD-3s with DB2. > > I am going to look into the Array Group activity monitor > > Another factor - DB@ I/O is low compared with our total IO profile > > Mike > > > On Tue, Dec 29, 2009 at 8:41 AM, Ron Hawkins > <ron.hawkins1...@sbcglobal.net>wrote: > > > Mike, > > > > > > > > What is "Cache Disk ratio?" > > > > > > > > Were you using PAV or HyperPAV on the smaller volumes where the DB2 > > datasets > > were previously allocated? > > > > > > > > I've benchmarked HDS storage from 9900 to USP-V and found no performance > > differences due to the volume size formatted on the array group > > (3390-3/9/27/54 and Custom volume Sizes). > > > > > > > > Have you checked the Array Group activity before and after the change to > > 3390-27? You can do this with your Performance Manager software launched > > from Storage Navigator. > > > > > > > > It's a common problem to end up with back-end IO spread over less spindles > > during volume consolidation, such that you end up with a sibling pend > > problem. Taking datasets spread over 50 or more array groups formatted as > > 3390-3 or 3390-9 and allocating them on just a few array groups formatted > > as > > 3390-27 will often cause the problem you are seeing. > > > > > > > > Ron > > > > > > > > > -----Original Message----- > > > > > From: IBM Mainframe Discussion List [mailto:ibm-m...@bama.ua.edu] On > > Behalf Of > > > > > Mike Shorkend > > > > > Sent: Monday, December 28, 2009 10:58 AM > > > > > To: IBM-MAIN@bama.ua.edu > > > > > Subject: [IBM-MAIN] HDS/USP Model 27 DASD and DB2 > > > > > > > > > > Hello All, > > > > > We have started experimenting with MOD-27 disks on a USP600 box. > > > > > We have found that with DB2, response times have deteriorated > > significantly. > > > > > We are seeing response times of more than 10ms on a fairly active disk. > > > > > Disconnect time accounts for about 90% of the time. Cache disk ratio has > > > > > gone down for these disks to less than 10%. > > > > > HYPERPAV is active - IOSQ time is zero. zHPF is not enabled. > > > > > > > > > > z/OS 1.9. DB2 V8. > > > > > > > > > > We have opened an issue with Hitachi. > > > > > Has anyone else experienced this in a similar configuration? > > > > > > > > > > Thanks > > > > > > > > > > Mike > > > > > > > > > > ---------------------------------------------------------------------- > > > > > For IBM-MAIN subscribe / signoff / archive access instructions, > > > > > send email to lists...@bama.ua.edu 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 lists...@bama.ua.edu with the message: GET IBM-MAIN INFO > > Search the archives at http://bama.ua.edu/archives/ibm-main.html > > > > > > -- > Mike Shorkend > m...@shorkend.com > www.shorkend.com > Tel: +972524208743 > Fax: +97239772196 > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@bama.ua.edu 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 lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
