Further more about this... In SC28-6823-02 (zSeries 890 and 990 CBU User's Guide) on page 1-1 it is mentioned that "Exceptions to this is the z890 where some CBU activations require a re-IPL." However, I have seen that in practice a re-IPL of z/OS is not suffient -- you must re-Activate the LPAR for the model identifier to change from x60 to x70. Any help will be appreciated, Gil.
On 5/20/05, Gil Peleg <[EMAIL PROTECTED]> wrote: > Hello all, > My question is about enabling the CBU feature on a z/890. > > We are running z/OS 1.4 on z/890 260 with 4 LPARS. Each LPAR is > defined in the activation profile with 2 initial CPs and 1 reserved. > After we enable the CBU feature, in the D M=CPU command I see our > machin is now a 370. > > Our initial intent was only to add a 3rd CP. But a change from 260 to > 370 not only implies the addition of a 3rd CP, but supposeably each CP > is now much stronger. > > I did not have a chance to fully examine the situation, but I have > noticed the following: > - when I use MXI to see how many MIPS the machine has before the CBU > is enabled it says 406, which is exactly what the capacity charts say > for 260. > - after enabling the CBU but before varying the 3rd CP online to the > LPAR, it still says 406 MIPS. > - after varying the 3rd CP online to the LPAR, is says 597 MIPS, which > is what the charts say for a 360 (not a 370!). > > I realize the MXI calculation is based on the RMCTADJC field in the > RMCT. I verified the value of RMCTADJC has not changed after of the > enablement of the CBU nor after varying the 3rd CP online to the LPAR. > > This would have made sense to me on any other machine. But on a z/890 > the strength of the CP can (supposably?) change instantly by the CBU > feature being enabled. > > I decided to further test the situation and re-IPLed a test LPAR with > 3 initial CPs. Then MXI noted 1045 MIPS, which is what the charts say > for a 370. Further more, after the IPL the RMCTADJC value was > different (smaller) then before the IPL. > > I couldnt find anything on the subject in the IBM Technical help database. > > I understand MXI is not a certified software i should rely on for such > a problem. What I wonder is: when running on a z/890, shouldn't z/OS > detect that the machine has changed from x60 to x70 and re-calculate > the RMCTADJC?? Otherwise, you end up with stronger CPs but it seems > z/OS does not realize it, so the whole service units calculation (for > example, for RMF reports) is wrong... > > Can anyone explain this behavior?? > > Thanks a lot, > Gil. > ---------------------------------------------------------------------- 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
