Setting Weights on LPARs

[Lizette Koehler]

I am trying to understand how weights work with LPARs.



We have 1 z890 with 2 LPARS (prod and test)



We set the weights to 65/35 for Prod vs. Test



Our DBAs said to raise it to 75/25 so that PROD would get more steam. We
only have 3 engines



They said when we did that it got better.



Does changing the weights from 65/35 to 75/25 in favor of PROD actually make
that much difference? This is not a hard cap. We are using the HMC to set
this up.





I have a couple of performance papers I have been reading (z9 Performance
Tuning Redbook). However, I am not seeing how this is significant.



Our TEST LPAR has the basic DB2, testing, verification, Qual check out and
major TSO and application development workload.



Our PROD LPAR has just the production work for DB2.



We know our real memory is tight so we added a little more to it this
weekend. Our DB2 environment in PROD is Data Sharing. Therefore, there is
another physical CPU (z9) with 3 LPARs (PROD, TEST, and SYSPROG)



The two PROD boxes each have Data Sharing members on them for our 2
Production DB2s.



We share dasd between all 5 LPARs.





Hypothesis: If I reset the weights on my 2 LPAR box to 60/40 or 75/25 what
would I see. What reports would help me see if this is doing any good? I
have MXG and SAS. I can also run RMFPP. I have not really gotten into
performance tuning on this level so I need to try and find the reports that
will help me understand this picture better.



Thanks

Lizette


----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@bama.ua.edu with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

[d***@ibm-main.lst]

Changing the weight from 65 to 75 (in your particular environment)
increased the available MIPS to the PROD LPAR by 15%: (75-65)/65=15.3%

Another simple example: if you have 1000 MIPS, your Prod LPAR went from
650 to 750 MIPS, which is not an insignificant change. Conversely, the
test LPAR went from 350 to 250 MIPS.

If you are not capping, both LPARs have access to MIPS that the other
LPAR is not using. The weights come into play when things get busy.

You can measure this by comparing things like TSO performance, batch
turnaround times, online response time, and by how loud the complaints
are (if any). MXG will give you info on all of that except for the
complaints. For those you're on your own.....

Dave Thorn * Senior Technology Analyst * SunGard Computer Services * 600
Laurel Oak Road, Voorhees, NJ, 08043
Tel [phone] * Mobile [phone] * Fax [phone]

CONFIDENTIALITY: This e-mail (including any attachments) may contain
confidential, proprietary and privileged information, and unauthorized
disclosure or use is prohibited. If you received this e-mail in error,
please notify the sender and delete this e-mail from your system.


-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-***@ibm-main.lst
Behalf Of Lizette Koehler
Sent: Friday, October 05, 2007 8:41 AM
To: IBM-***@BAMA.UA.EDU
Subject: Setting Weights on LPARs

Hypothesis: If I reset the weights on my 2 LPAR box to 60/40 or 75/25
what
would I see. What reports would help me see if this is doing any good?
I
have MXG and SAS. I can also run RMFPP. I have not really gotten into
performance tuning on this level so I need to try and find the reports
that
will help me understand this picture better.


Thanks

Lizette


----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@bama.ua.edu with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@bama.ua.edu with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

[Mark Zelden]

On Fri, 5 Oct 2007 08:40:37 -0400, Lizette Koehler <***@MINDSPRING.COM>
wrote:


>We have 1 z890 with 2 LPARS (prod and test)
>We set the weights to 65/35 for Prod vs. Test
>Our DBAs said to raise it to 75/25 so that PROD would get more steam. We
>only have 3 engines
>They said when we did that it got better.

>Does changing the weights from 65/35 to 75/25 in favor of PROD actually make
>that much difference? This is not a hard cap. We are using the HMC to set
>this up.
>
>

Of course it can... if the box is running at or near capacity. It also depends
how much CPU queuing was going on. "A little" extra could get you over
the hump.

How many engines on test? If only 1, then anything over 33 could not be
achieved anyway.

Mark
--
Mark Zelden
Sr. Software and Systems Architect - z/OS Team Lead
Zurich North America / Farmers Insurance Group - ZFUS G-ITO
mailto:***@zurichna.com
z/OS and OS390 expert at http://searchDataCenter.com/ateExperts/
Systems Programming expert at http://expertanswercenter.techtarget.com/
Mark's MVS Utilities: http://home.flash.net/~mzelden/mvsutil.html

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@bama.ua.edu with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

[Joel Wolpert]

The weights only take effect when the machine is at 100%. You should check
the CPU utilization of the machine and the lpar in RMF. If the machine is
at 100% and the lpar is at 100% than this most probably did help out.

Lizette said:

I am trying to understand how weights work with LPARs.



We have 1 z890 with 2 LPARS (prod and test)



We set the weights to 65/35 for Prod vs. Test



Our DBAs said to raise it to 75/25 so that PROD would get more steam. We
only have 3 engines



They said when we did that it got better.



Does changing the weights from 65/35 to 75/25 in favor of PROD actually
make
that much difference? This is not a hard cap. We are using the HMC to set
this up.





I have a couple of performance papers I have been reading (z9 Performance
Tuning Redbook). However, I am not seeing how this is significant.



Our TEST LPAR has the basic DB2, testing, verification, Qual check out and
major TSO and application development workload.



Our PROD LPAR has just the production work for DB2.



We know our real memory is tight so we added a little more to it this
weekend. Our DB2 environment in PROD is Data Sharing. Therefore, there is
another physical CPU (z9) with 3 LPARs (PROD, TEST, and SYSPROG)



The two PROD boxes each have Data Sharing members on them for our 2
Production DB2s.



We share dasd between all 5 LPARs.





Hypothesis: If I reset the weights on my 2 LPAR box to 60/40 or 75/25
what
would I see. What reports would help me see if this is doing any good? I
have MXG and SAS. I can also run RMFPP. I have not really gotten into
performance tuning on this level so I need to try and find the reports that
will help me understand this picture better.



Thanks

Lizette



Joel Wolpert
Director - Performance and Capacity Planning
Shared Data Center
Securities Industry Automation Corporation
2 Metrotech Center
New York, NY 11201
[phone]
***@siac.com





Visit our website at http://www.nyse.com

****************************************************

Note: The information contained in this message and any attachment
to it is privileged, confidential and protected from disclosure. If the
reader of this message is not the intended recipient, or an employee
or agent responsible for delivering this message to the intended
recipient, you are hereby notified that any dissemination,
distribution or copying of this communication is strictly prohibited.
If you have received this communication in error, please notify the
sender immediately by replying to the message, and please delete
it from your system. Thank you. NYSE Group, Inc.

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@bama.ua.edu with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

[Mark Zelden]

Herbie,

I monitor IBM-MAIN via the web interface and I haven't seen
your post yet. Maybe it was rejected due to the images you
attached (deleted from my response). To answer your question...

The weights are normalized to 100%. So 33, is the same as
333 if using 1000 as the total instead of 100. It is common to
use 1000 since it allows more granularity.

However, in the screen shots you attached, the total of the
weights is 1200 (well, 1199). So 200 /1200 is 16%, which
is still less than the max of 33% of the box if you have one
engine defined (which means it is achievable).

The production LPAR has 84%. So it better have 3 enginges
defined to it. If it only had 2, it could only get up
to 66% of the total capacity of the box - even if the test
lpar was only using 1%.

Mark
--
Mark Zelden
Sr. Software and Systems Architect
Zurich North America / Farmers Insurance Group - ZFUS G-ITO
mailto:***@zurichna.com
z/OS and OS390 expert at http://searchDataCenter.com/ateExperts/com/
Systems Programming expert at http://expertanswercenter.techtarget.com/
Mark's MVS Utilities: http://home.flash.net/~mzelden/mvsutil.html





From: Van Dalsen, Herbie
Sent: 05 Oktober 2007 02:38 nm
To: 'IBM Mainframe Discussion List'
Subject: RE: Setting Weights on LPARs

Mark,

We also have a z890, but came from a H30... AS will Richard well know....
On the Customize Activation Profiles it allows you to go to 999 ? what on
earth for if what you say is true about the 33. AS you can see, we run our
dev/prod with 999 / 200, I suppose 9,99 / 2 ?

Regards

Herbie





******************* PLEASE NOTE *******************
This E-Mail/telefax message and any documents accompanying this
transmission may contain privileged and/or confidential information and is
intended solely for the addressee(s) named above. If you are not the
intended addressee/recipient, you are hereby notified that any use of,
disclosure, copying, distribution, or reliance on the contents of this
E-Mail/telefax information is strictly prohibited and may result in legal
action against you. Please reply to the sender advising of the error in
transmission and immediately delete/destroy the message and any
accompanying documents. Thank you.

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@bama.ua.edu with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

[Ted MacNEIL]

>The weights only take effect when the machine is at 100%.

Not quite accurate.

All LPARs also have to 'want' CPU.
If one is coasting and the other wants more, it will get it (unless capped).

EG:
Weights: LPARA -- 650
LPARB -- 350

Off shift LPARA is nearly idle
Processor usage 100%

From real life:
Usage: LPARA -- 30%
LPARB -- 69%
PHYSICAL -- 1%

Weights only take effect if there is processor contention (at 100%).

-
Too busy driving to stop for gas!

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to ***@bama.ua.edu with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html