I break it into two parts:  hardware-motivated size changes and 
software-motivated ones.  Normally the latter are fairly predictable, e.g. 
based on usage, etc.  Here you say there is no indication of any change there.

The former are somewhat sparsely documented in my view, but there is usually 
mention in the PR/SM guide and the various redbooks (Tech Guide, Tech Intro, 
etc.) of what a new level of CFCC will mean for the CF structure sizes.  The 
z14 Tech Intro says this in section 3.7.6:

 Coupling Facility Control Code Level
Various levels (such as 22 and 23) of Coupling Facility Control Code (CFCC) are 
available for the z14. For more information, see IBM Knowledge Center.
CF structure size changes can be expected when moving from one CFCC level to 
another.  Review the CF LPAR size by using the following tools:
 The CFSizer tool is a web-based and is useful when a workload is changed or 
introduced
 The Sizer Utility, which is an authorized z/OS program download, is useful 
when upgrading a CF

Have you tried the Sizer Utility?  I don't have a system to use it on just now 
but it may be worth a shot.  I also found that the CFSizer tool can be a lot of 
work to use and would avoid it if possible  :-)

A ROT:  aside from your usual best-practice monitoring of changes due to usage, 
only change the size if the CFCC code needs it.  After using CFSizer and the 
various inputs it needed for the different structures, I seem to recall having 
to approximately double the GRS structure sizes when we went to z13 and level 
21, and that that ended up being overkill although we left them in place as 
they weren't impacting anything.  DB2 structures and others (RACF, VSAM RLS, 
etc.) were unchanged.  I didn't know about the Sizer Utility then, it may have 
saved us the hassle of making the GRS changes.

These days with "cheap" memory it can pay to oversize the CF LPARs for safety's 
sake.  Also, if you have a sysprog-only Sysplex, which I assume you do as that 
is also best practice, then one can consider resizing CFs with new LPAR 
definitions, structure moves, etc. if it is worth doing so after the initial 
migration is bedded in.  I tend to take a cautious approach to CF operation, as 
I believe you do going by your previous posts.

cheers,
Peter


On Thu, 25 Oct 2018 22:46:13 +0000, Jesse 1 Robinson <[email protected]> 
wrote:

>We've played the game of sizing CF structures for two decades. About to play 
>it yet again. The main problem is that the sizing tool asks for some input for 
>most structures, and that input is itself a SWAG. Without spending an 
>inordinate amount of time researching the current environment, does anyone 
>have a ROT for estimating structure sizes for a new CEC? Most structures are 
>probably OK-they usually are-but some may need to increase, an exigency that 
>often does not become apparent until a system gets busy on a Tuesday morning.
>
>We're moving from two z12 CECs to a z14 and a z13s with no sizing problems 
>currently in evidence. Any advice?
>
>
>.
>.
>J.O.Skip Robinson
>Southern California Edison Company
>Electric Dragon Team Paddler
>SHARE MVS Program Co-Manager
>323-715-0595 Mobile
>626-543-6132 Office <===== NEW
>[email protected]<mailto:[email protected]>
>
>
>----------------------------------------------------------------------
>For IBM-MAIN subscribe / signoff / archive access instructions,
>send email to [email protected] with the message: INFO IBM-MAIN

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to