R.S. wrote:
W dniu 2010-08-21 17:18, Steve Comstock pisze:
R.S. wrote:
W dniu 2010-08-21 15:30, Steve Comstock pisze:
Peter Relson wrote:
Regardless of how many CPs an LPAR may allow there may be an operating
system restriction to a smaller number.
It happens that at this point in time z/OS (1.11) is a bit ahead of
the LPAR limit so that it can support what is allowed within the LPAR.
I don't know about other operating systems.
Hmmm. So, can z/OS support 80 x 32 = 2560 active CPs per sysplex?
Assuming you need the number to some slide slideshow, just put
"approx. 3000 CP's". Fine print for "approx." ;-)
It sound quite impressive. Details are irrelevant.
Well, it is for a class, and not for managers, for techies,
and they want to know actual acheivable limits, and I think
that the only correct way to do that is precisely.
I dare to disagree. While it is meant for techies, it's not technical,
rather theoretical. It's like considering building constructions on
Venus or memory above 4GB for Novell Netware in 1990. In simple words
nobody will buy 32 or 48 z196s. Such customer could get dedicated
version of both hardware and z/OS, with some limitations relieved.
(Rhetorical:) Would it be reasonable to create OS/390 V2R6 cabaple to
run on 100 CPs while the largest machine had no more than 12 and there
were plan s for 16CPs?
Well, that's a good point. I remember a few years back someone
posted on this group that if you thought about 64-bit address
spaces it's really absurd:
1. The system could not support enough DASD capacity to
fully page-out a single 16 EB address space, and besides
the maximum number of page data sets limits you even
further
2. If you actually did have to page in a 64-bit address space
at current data transfer rates it would take around 500
years (or was it 500 hours? my memory is a little hazy
on their math; either way it's untenable)
Point being your point: one may have theoretical limits
and practical limits - which is important?
Havine the above in mind IMHO it's quite OK to say "the limit is 80x32,
maybe some other factors coulds slightly lower it." Details can be left
for real "due diligence" project.
[...]
Sure. But what is acheivable as delivered if you ordered
32 of the largest z196's?
Not 32. Add 16 for CFs. You don't want to lower the limit of available
CPs. ;-)
Why 16 CFs? Because it is architectural limit! ;-)
Does the limit make any sense? Yes, of course: it's MORE THAN YOU NEED.
No one will ever need more than 640K
5 mainframes
16 MB
Probably some other limitations like # of CF links or former # of timer
links can reduce this theoretical limit.
Regards
--
Kind regards,
-Steve Comstock
The Trainer's Friend, Inc.
303-393-8716
http://www.trainersfriend.com
* To get a good Return on your Investment, first make an investment!
+ Training your people is an excellent investment
----------------------------------------------------------------------
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
