I stand corrected. We must have been on V8 long enough for V7 "64-bit-backed" to mind-morph into "64-bit-addressable". The net effect as you mention was that by using one or more dataspaces for buffer pools, DB2 V7 was no longer bound by the 2GiB limit for total virtual space or total real storage utilization. JC Ewing
On 02/05/2010 05:36 PM, Wayne Driscoll wrote: > Dataspaces (and hiperspaces, which is actually what DB2 used for > bufferpool extensions prior to DB2 V8) have a maximum size of 2 GiB even > in z/OS so it is correct that DB2 V7 was strictly 31 bit. The use of > hiperpools and EDMPOOL in a dataspace allowed DB2 virtually storage to > grow "horizontally" by having additional data or hiper spaces each a max > of 2GiB in size. > > > =============================================== > Wayne Driscoll > OMEGAMON DB2 L3 Support/Development > wdrisco(AT)us.ibm.com > =============================================== > > > > From: > "Joel C. Ewing" <jcew...@acm.org> > To: > IBM-MAIN@bama.ua.edu > Date: > 02/05/2010 05:23 PM > Subject: > Re: DB2 V7 - Implications to storage increasing RID Pool > Sent by: > IBM Mainframe Discussion List <IBM-MAIN@bama.ua.edu> > > > > This is not completely true. Not everything in DB2 V7 was required to > be under the bar. Although the main address space was only 31-bit, V7 > supported placing buffer pools in a 64-bit-addressable dataspace so that > DB2 V7 could effectively make use of more than 2 GB of real storage. We > ran in that mode for a number of years before going to V8. > JC Ewing > > On 02/03/2010 04:38 PM, Mohammad Khan wrote: >> V7 is a 31-bit app therefore everything is under the bar there. Increase > the >> pool only if you have enough real memory to go with it otherwise results > may >> not be to your liking. Even if there is enough real memory to go with it > do >> check that you are not taking away virtual storage from another critical > part >> of DBM1 address space e.g. virtual buffer pools, remember it all has to > fit in >> under 2GB. This increase can help with DB2 performance if you are >> encountering many RID pool failures DUE TO POOL SHORTAGE but doesn't > help >> if the failure is due RDS RID limit. Then again you may be able to tune > some of >> these queries to use better access path and not rely on RID soting. >> HTH >> Mohammad >> >> >> On Wed, 3 Feb 2010 04:06:55 -0800, Patrick Falcone >> <patrick.falco...@verizon.net> wrote: >> >>> Curious if anyone can shed light on the subject matter. We're taking > the >> default of 4 mb and want to increase to around 100 mb for the RID pool. > It >> appears in V7 the RID is allocated below the bar, in V8 it looks like > the lists >> part of the RID get moved above the bar. >>> >>> Do we need to be concerned about region or any other storage metrics, >> performance implications to DB2? >>> >>> TIA... >>> > > -- Joel C. Ewing, Fort Smith, AR jremoveccapsew...@acm.org ---------------------------------------------------------------------- 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
