>PR/SM can use additional engines
>lowering the logical to physical ratio, but MVS only builds enough control
>blocks on for the number of initial engines plus the number of reserved
>engines.
>This may have been improved with recent releases of z/OS but this was
>certainly an issue in the past.
As Mark Zelden has correctly commented, z/OS 1.10 supports the z10
functionality that lets you define to an LPAR a new CPU that was not
reserved and then you can configure it online to z/OS without an IPL. As a
result there are some things you might notice, such as on such a machine
CVTMAXMP not being the actual highest CPU number known at IPL but instead
being the highest CPU number possible for that machine (and, similarly, the
LCCA and PCCA vector tables being of maximum-possible size). There is a new
field that could be of interest (and you might consider changing
applications to use it instead of CVTMAXMP):
ECVT_Installed_CPU_HWM DC H'0' The highest CPU number currently
installed within this IPL. Could
increase upon dynamic CPU addition
Applications that allocate storage per-possible-processor and then depend
on that storage being available whenever they find themselves running on
"CPU n", basing that storage allocation upon the value in CVTMAXMP is
usually the right thing to do.. But applications that simply are
"reporting" by looping through possible processors might well do fine using
ECVT_Installed_CPU_HWM.
Peter Relson
z/OS Core Technology Design
----------------------------------------------------------------------
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
