Re: CPU Contention Issue
- From: mikey.moss@xxxxxxxxxxxx (Michael W. Moss)
- Date: 2 Jan 2006 13:00:00 -0800
Hi Jacky,
The information you ask for regarding Peter Enrico might be available here
in the form of Technical Presentations and many other links:
http://www.schunk-associates.com/
There is also a whole lot of information at the IBM web site, so just
search for Peter?s name and/or WLM
With regard to a CPU running at 100%, then potentially this is what one
might want to do, run the box at capacity to get ?maximum bang for your
buck?. However, generally we will find that our ?Workload? suffers when
this is the case. I?m not so sure that upgrading the CPU is a technology
issue, but certainly it's a business issue. So at the really basic level
this is your business meeting their Service Level Agreements (SLA?s), for
example:
* 99.5% of CICS transactions complete in less than 1 second
* Critical Path Batch Processing completes from 00:00 to 06:00
* DR vaulting completes by 06:00 every day
These are really simplistic and pretty crude, but ultimately some business
applications might be more critical than others, and WLM goals should meet
such SLA?s accordingly. Maybe you have some rogue processes that are
consuming resource, or maybe there might be performance advantages in
upgrading to 64-bit if you haven?t already done so, and the latest
versions of software in general.
Bottom Line: Ultimately your business should ?tell you? when the time for
a CPU upgrade is when SLA?s aren?t being delivered. Optionally a CPU
upgrade might be performed to lower the TCO; so upgrading to a latest
generation box might actually reduce TCO over a several year period,
generating tangible cost ROI, while generally delivering those tangible
business benefits, which can only be described in ?Plain English? via an
SLA and translated into ?IBM Mainframe Speak? by WLM
Regards, MWM
On Sun, 1 Jan 2006 16:18:01 +0530, Jacky Bright <jacky.bright@xxxxxxxxx>
wrote:
>This is with reference to RMF Report analysis regarding MVS Busy and LPAR
>Busy parameter ...
>
>Attached is the RMF Report.
>
>Almost everyday we are facing CPU Contetion issues.
>
>>From the report we can infer tht for most of the intervals there were 100%
>CPU Contention however is it possible to get exact reason for the CPU
>Contenction from RMF Report ??
>
>Wht could be the reasons for CPU contention ?
>
>Are there any performance related parameters which need to consider while
>analysing these RMF reports ??
>
>I heard some MVS performance parameters like ratio of LPAR to MVS Busy (by
>Peter Enrico) can anyone send me any reference material (by Peter Enrico)
in
>tht regard ??
>
>For some intervals ther is MVS Busy is 100 % where as LPAR Busy is < 100%.
>Can we say tht thr is requirement of upgradation of CPU on the basis of
>these RMF Reports ??
>
>We have 2 CICS regions out of which one CICS region uses C++ programmin
>while other uses COBOL programmin recently it has been observed tht COBOL
>CICS region is hoggin the CPU and which in turn affects C++ CICS region.
>
>Under what conditions we can say tht there is requirement of CPU
Upgradation
>??
>
>what should be the policy of organisation for S/390 CPU Upgradation ?
>
>
>
>Jacky
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to listserv@xxxxxxxxxxx with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
.
- Prev by Date: CSA page tables
- Next by Date: PDF? (was: When was the leap second inserted?)
- Previous by thread: CPU Contention Issue
- Next by thread: Re: EREP , sense ... manual
- Index(es):
Relevant Pages
|
