Understanding the LPAR's maximum limit.

How to do Performance analysis and Capacity Planning for a shop. We also talk about Software Cost here.
Post Reply
VajraC
New Member
Posts: 2
Joined: Sun Nov 23, 2014 1:40 pm

Understanding the LPAR's maximum limit.

Post by VajraC »

Hi,

Let's assume we have a box with 700 MSU and there are 4 LAPRs on this box. Now LPAR-1 is assigned 4 logical processors & LPAR 2,3 and 4 are assigned 1 logical processor each. So tatal of 7 logical processors.


Now how much maximum MSU can LPAR 1 can run with, is it 400? And LPAR 2, 3 and 4 run at maximum of 100 MSU's each? Please guide.
User avatar
Robert Sample
Global Moderator
Global Moderator
Posts: 1891
Joined: Fri Jun 28, 2013 1:22 am
Location: Dubuque Iowa
United States of America

Re: Understanding the LPAR's maximum limit.

Post by Robert Sample »

How many physical processors make up the box? What are the weights of the LPARs? I think your question is confusing -- "logical processor" is an LPAR (logical partition) while physical processor is the number of hardware processors on the machine. And without knowing the LPAR weights (from the IODF) you cannot really answer the question. Physical processors can be CP (central processors), zAAP, zIIP, ICF and they are treated differently in the IODF.

I will assume you meant there are 4 physical processors in the CEC (that is, the hardware box). Furthermore, this machine has only CP (unlikely I know but it simplifies things), and that there are 4 LPARs defined with weights of 700, 100, 100, 100 respectively. The 700 MSU is split evenly between the 4 physical processors so each processor has 175 MSU. If the 4 LPARs are allowed to each use all four physical processors, and capping is not being used, then the MSU ratings for the 4 LPARs would be 490, 70, 70, 70 respectively (700 / 1000 weight times 700 MSU, 100 / 1000 times 700 MSU, 100 / 1000 times 700 MSU, 100 / 1000 times 700 MSU). If the physical processors are not evenly split between the 4 LPARs, then the MSU weights would be different. A spreadsheet could be used to calculate each processor usage by each LPAR under different assumptions.

Share usually has a presentation or two on LPAR management, or you can Google z/OS lpar weight for more details.
VajraC
New Member
Posts: 2
Joined: Sun Nov 23, 2014 1:40 pm

Re: Understanding the LPAR's maximum limit.

Post by VajraC »

Sorry, I meant physical processors as you have correctly said.

If the physical processors are not evenly split between the 4 LPARs, then the MSU weights sould be different as you said but then how do one decides on how to distribute the processors, to start with?
User avatar
Robert Sample
Global Moderator
Global Moderator
Posts: 1891
Joined: Fri Jun 28, 2013 1:22 am
Location: Dubuque Iowa
United States of America

Re: Understanding the LPAR's maximum limit.

Post by Robert Sample »

When an upgrade is being done, the current configuration is used to determine how to allocate processing in the new configuration. Generally a study will be done to verify that the new configuration LPAR weights will adequately handle the workload. Frequently, experienced capacity planners will tweak the results to match their expectations of the system and its performance. An LPAR is generally used for production, test, development, or sandbox (operating systems and products upgrades) -- larger sites may have multiple LPARs dedicated to some of these functions -- and site management ultimately determines the relative weights (after considering the results of the configuration study and the capacity analyst's recommendations).

The SMF type 70 records can be used to identify the LPARs and their assigned weights. There is also information about the CEC in the SMF type 70 records.
Post Reply

Create an account or sign in to join the discussion

You need to be a member in order to post a reply

Create an account

Not a member? register to join our community
Members can start their own topics & subscribe to topics
It’s free and only takes a minute

Register

Sign in

Return to “Mainframe Performance, Capacity Planning and Software Cost.”