How to identify what is process is creating Unaccounted time in DB2

Lizette Koehler

How to identify what is process is creating Unaccounted time in DB2
I am trying to understand why a DB2 application runs fine when only one user is
running, but gets really slow when it gets up to 50. The applications team
feels they should be able to handle several hundred users at one time.



We are using DB2 PM and Detector to do the analysis



Looking for guidance on isolating and identifying what is "unaccounted time" in
DB2 as identified by CA Detector



I understand it could be as simple as waiting on HSM recall. But there were no
recalls when the users were executing their DB2 application.



And my very basic understanding is that DB2 is "waiting" for something to
complete.



What we are seeing is a large part of the DB2 work seems to be in something
called "unaccounted" time.



When we look at RMF, the CPU is not heavily used (less than 50% busy). RMF
shows DB2 waiting on CPU. There is no paging going on.

No error messages from CF or HSM or other subsystems on the LPAR.



This is from a server to DB2 then back, and no messages that seem to indicate
any issues within z/OS



Could someone help me understand what is "unaccounted time" is from DB2 that
would be and how I would go about determining where this process is doing in DB2
when that this occurs. How can I identify what it is DB2 is requesting that is
causing it to wait.



I will be doing more research on the internet on this, but thought someone could
provide me with some pointers on this area.





Thank you



Lizette Koehler

statistics: A precise and logical method for stating a half-truth inaccurately



Phil Grainger

How to identify what is process is creating Unaccounted time in DB2
(in response to Lizette Koehler)
If I remember rightly, Accounting Class 3 if started gives more granularity on "unknown" wait timers (just checked the KC "Accounting class 3 data provides detailed information about the distribution of suspension times and related events.")

So I assume you are currently running without Class 3?

Also see https://www.ibm.com/support/knowledgecenter/en/SSEPEK_11.0.0/perf/src/tpc/db2z_investigatehighclass3.html

Phil Grainger
Principal Enablement Manager

[BMC Exchange 2019 - Global Event Series - REGISTER] https://www.bmc.com/ami

Direct

+44 1189 218 000

Mobile

+44 7808 643 479

Email

[login to unmask email]

E2, Eskdale Road
Winnersh
Berkshire
United Kingdom
RG41 5TS
[image001 (002)][https://acclaim-production-app.s3.amazonaws.com/images/2429c3cd-a1de-44fc-b4f3-bc762bb2f963/IBM%2BChampion%2B-%2BAnalytics%2B2018.png][https://acclaim-production-app.s3.amazonaws.com/images/2429c3cd-a1de-44fc-b4f3-bc762bb2f963/IBM%2BChampion%2B-%2BAnalytics%2B2018.png]



From: Lizette Koehler <[login to unmask email]>
Sent: 10 July 2019 13:48
To: [login to unmask email]
Subject: [EXTERNAL] [DB2-L] - How to identify what is process is creating Unaccounted time in DB2

I am trying to understand why a DB2 application runs fine when only one user is running, but gets really slow when it gets up to 50. The applications team feels they should be able to handle several hundred users at one time.

We are using DB2 PM and Detector to do the analysis

Looking for guidance on isolating and identifying what is "unaccounted time" in DB2 as identified by CA Detector

I understand it could be as simple as waiting on HSM recall. But there were no recalls when the users were executing their DB2 application.

And my very basic understanding is that DB2 is "waiting" for something to complete.

What we are seeing is a large part of the DB2 work seems to be in something called "unaccounted" time.

When we look at RMF, the CPU is not heavily used (less than 50% busy). RMF shows DB2 waiting on CPU. There is no paging going on.
No error messages from CF or HSM or other subsystems on the LPAR.

This is from a server to DB2 then back, and no messages that seem to indicate any issues within z/OS

Could someone help me understand what is "unaccounted time" is from DB2 that would be and how I would go about determining where this process is doing in DB2 when that this occurs. How can I identify what it is DB2 is requesting that is causing it to wait.

I will be doing more research on the internet on this, but thought someone could provide me with some pointers on this area.


Thank you

Lizette Koehler
statistics: A precise and logical method for stating a half-truth inaccurately


-----End Original Message-----
BMC Software Limited Registered Office: Building E2, Eskdale Road, Winnersh, Wokingham, Berkshire, United Kingdom, RG41 5TS Registered in England No. 1927903 The content of this email is confidential. If you are not the addressee, you may not distribute, copy or disclose any part of it. If you receive this message in error, please delete this from your system and notify the sender immediately.
Attachments

  • image001.jpg (49.7k)
  • image002.png (6.7k)
  • image003.png (3.7k)
  • image004.png (<1k)

John Bucaria

How to identify what is process is creating Unaccounted time in DB2
(in response to Lizette Koehler)
It has been my understanding that the Detector "UNACC_TIME" generally referred to CPU wait but nothing more granular.
-John

From: Lizette Koehler <[login to unmask email]>
Sent: Wednesday, July 10, 2019 8:48 AM
To: [login to unmask email]
Subject: [DB2-L] - How to identify what is process is creating Unaccounted time in DB2

I am trying to understand why a DB2 application runs fine when only one user is running, but gets really slow when it gets up to 50. The applications team feels they should be able to handle several hundred users at one time.

We are using DB2 PM and Detector to do the analysis

Looking for guidance on isolating and identifying what is "unaccounted time" in DB2 as identified by CA Detector

I understand it could be as simple as waiting on HSM recall. But there were no recalls when the users were executing their DB2 application.

And my very basic understanding is that DB2 is "waiting" for something to complete.

What we are seeing is a large part of the DB2 work seems to be in something called "unaccounted" time.

When we look at RMF, the CPU is not heavily used (less than 50% busy). RMF shows DB2 waiting on CPU. There is no paging going on.
No error messages from CF or HSM or other subsystems on the LPAR.

This is from a server to DB2 then back, and no messages that seem to indicate any issues within z/OS

Could someone help me understand what is "unaccounted time" is from DB2 that would be and how I would go about determining where this process is doing in DB2 when that this occurs. How can I identify what it is DB2 is requesting that is causing it to wait.

I will be doing more research on the internet on this, but thought someone could provide me with some pointers on this area.


Thank you

Lizette Koehler
statistics: A precise and logical method for stating a half-truth inaccurately


-----End Original Message-----

Joel Goldstein

How to identify what is process is creating Unaccounted time in DB2
(in response to John Bucaria)
If DB2 is "waiting on cpu" - the cpu is busy servicing some other
application or function. What besides DB2 are the heaviest cpu users?



Your processor is only 50% busy ---



Is there more than one engine available to this lpar?



Are cpu engines shared over multiple lpars?



IMO - look to your WLM definitions









Joel Goldstein
Responsive Systems
Buffer Pool Tool(R) for DB2, the worldwide industry standard

Predicts the IO rate/Sec for tuning changes
Performance software that works......
Predicts Group Buffer Pool performance too!
http://www.responsivesystems.com www.responsivesystems.com
tel. (732) 972-1261
fax.(732) 972-9416

[login to unmask email]



From: Bucaria, John [mailto:[login to unmask email]
Sent: Wednesday, July 10, 2019 7:07 AM
To: [login to unmask email]
Subject: [DB2-L] - RE: How to identify what is process is creating
Unaccounted time in DB2



It has been my understanding that the Detector "UNACC_TIME" generally
referred to CPU wait but nothing more granular.

-John



From: Lizette Koehler <[login to unmask email]>
Sent: Wednesday, July 10, 2019 8:48 AM
To: [login to unmask email]
Subject: [DB2-L] - How to identify what is process is creating Unaccounted
time in DB2



I am trying to understand why a DB2 application runs fine when only one user
is running, but gets really slow when it gets up to 50. The applications
team feels they should be able to handle several hundred users at one time.



We are using DB2 PM and Detector to do the analysis



Looking for guidance on isolating and identifying what is "unaccounted time"
in DB2 as identified by CA Detector



I understand it could be as simple as waiting on HSM recall. But there were
no recalls when the users were executing their DB2 application.



And my very basic understanding is that DB2 is "waiting" for something to
complete.



What we are seeing is a large part of the DB2 work seems to be in something
called "unaccounted" time.



When we look at RMF, the CPU is not heavily used (less than 50% busy). RMF
shows DB2 waiting on CPU. There is no paging going on.

No error messages from CF or HSM or other subsystems on the LPAR.



This is from a server to DB2 then back, and no messages that seem to
indicate any issues within z/OS



Could someone help me understand what is "unaccounted time" is from DB2 that
would be and how I would go about determining where this process is doing in
DB2 when that this occurs. How can I identify what it is DB2 is requesting
that is causing it to wait.



I will be doing more research on the internet on this, but thought someone
could provide me with some pointers on this area.





Thank you



Lizette Koehler

statistics: A precise and logical method for stating a half-truth
inaccurately





-----End Original Message-----



-----End Original Message-----

Lizette Koehler

How to identify what is process is creating Unaccounted time in DB2
(in response to John Bucaria)
Thanks.


We do have Trace 3 running (Accounting) for 01,02,03 going to SMF



Lizette







From: Bucaria, John <[login to unmask email]>
Sent: Wednesday, July 10, 2019 7:07 AM
To: [login to unmask email]
Subject: [DB2-L] - RE: How to identify what is process is creating Unaccounted
time in DB2



It has been my understanding that the Detector "UNACC_TIME" generally referred
to CPU wait but nothing more granular.

-John



From: Lizette Koehler <[login to unmask email] <mailto:[login to unmask email]> >
Sent: Wednesday, July 10, 2019 8:48 AM
To: [login to unmask email] <mailto:[login to unmask email]>
Subject: [DB2-L] - How to identify what is process is creating Unaccounted time
in DB2



I am trying to understand why a DB2 application runs fine when only one user is
running, but gets really slow when it gets up to 50. The applications team
feels they should be able to handle several hundred users at one time.



We are using DB2 PM and Detector to do the analysis



Looking for guidance on isolating and identifying what is "unaccounted time" in
DB2 as identified by CA Detector



I understand it could be as simple as waiting on HSM recall. But there were no
recalls when the users were executing their DB2 application.



And my very basic understanding is that DB2 is "waiting" for something to
complete.



What we are seeing is a large part of the DB2 work seems to be in something
called "unaccounted" time.



When we look at RMF, the CPU is not heavily used (less than 50% busy). RMF
shows DB2 waiting on CPU. There is no paging going on.

No error messages from CF or HSM or other subsystems on the LPAR.



This is from a server to DB2 then back, and no messages that seem to indicate
any issues within z/OS



Could someone help me understand what is "unaccounted time" is from DB2 that
would be and how I would go about determining where this process is doing in DB2
when that this occurs. How can I identify what it is DB2 is requesting that is
causing it to wait.



I will be doing more research on the internet on this, but thought someone could
provide me with some pointers on this area.





Thank you



Lizette Koehler

statistics: A precise and logical method for stating a half-truth inaccurately





-----End Original Message-----



-----End Original Message-----

Fred Edgar

How to identify what is process is creating Unaccounted time in DB2
(in response to Joel Goldstein)
This link lists the possible things responsible for UNACC time:  IBM What is DB2 Accounting Class 2 Not Accounted Time?

|
|
| |
IBM What is DB2 Accounting Class 2 Not Accounted Time?

What is DB2 Accounting Class 2 Not Accounted Time?
|

|

|




On Wednesday, July 10, 2019, 09:42:22 AM CDT, Joel Goldstein - Responsive Systems <[login to unmask email]> wrote:


If DB2 is “waiting on cpu” – the cpu is busy servicing some other application or function.  What besides DB2 are the heaviest cpu users?

 

Your processor is only 50% busy ---  

 

Is there more than one engine available to this lpar?

 

Are cpu engines shared over  multiple lpars?

 

IMO -  look to your WLM definitions

 

 

 

 

Joel Goldstein
Responsive Systems
Buffer Pool Tool(R) for DB2, the worldwide industry standard

         Predicts the IO rate/Sec for tuning changes
         Performance software that works......
   Predicts Group Buffer Pool performance too!
www.responsivesystems.com
tel. (732) 972-1261
fax.(732) 972-9416

[login to unmask email]

 

From: Bucaria, John [mailto:[login to unmask email]
Sent: Wednesday, July 10, 2019 7:07 AM
To: [login to unmask email]
Subject: [DB2-L] - RE: How to identify what is process is creating Unaccounted time in DB2

 

It has been my understanding that the Detector “UNACC_TIME” generally referred to CPU wait but nothing more granular.

-John

 

From: Lizette Koehler <[login to unmask email]>
Sent: Wednesday, July 10, 2019 8:48 AM
To: [login to unmask email]
Subject: [DB2-L] - How to identify what is process is creating Unaccounted time in DB2

 

I am trying to understand why a DB2 application runs fine when only one user is running, but gets really slow when it gets up to 50.  The applications team feels they should be able to handle several hundred users at one time.

 

We are using DB2 PM and Detector to do the analysis

 

Looking for guidance on isolating and identifying what is “unaccounted time” in DB2 as identified by CA Detector

 

I understand it could be as simple as waiting on HSM recall.  But there were no recalls when the users were executing their DB2 application.

 

And my very basic understanding is that DB2 is “waiting” for something to complete.

 

What we are seeing is a large part of the DB2 work seems to be in something called “unaccounted” time.

 

When we look at RMF, the CPU is not heavily used (less than 50% busy).  RMF shows DB2 waiting on CPU.  There is no paging going on.

No error messages from CF or HSM or other subsystems on the LPAR.

 

This is from a server to DB2 then back, and no messages that seem to indicate any issues within z/OS

 

Could someone help me understand what is “unaccounted time” is from DB2 that would be and how I would go about determining where this process is doing in DB2 when that this occurs.  How can I identify what it is DB2 is requesting that is causing it to wait.

 

I will be doing more research on the internet on this, but thought someone could provide me with some pointers on this area.

 

 

Thank you

 

Lizette Koehler

statistics: A precise and logical method for stating a half-truth inaccurately

 

 

-----End Original Message-----

 
-----End Original Message-----
Site Links: View post online   View mailing list online   Start new thread via email   Unsubscribe from this mailing list   Manage your subscription  

This email has been sent to: [login to unmask email]
Discover the best cloning tool on the market. Try BCV5 & the new BCV5 Masking Tool.
ESAi also has powerful solutions for Buffer Pool Tuning, Log Analysis, TDM, & more.
http://www.ESAIGroup.com/idug



Use of this email content is governed by the terms of service at:
http://www.idug.org/p/cm/ld/fid=2

Philip Sevetson

How to identify what is process is creating Unaccounted time in DB2
(in response to Phil Grainger)
Phil G.,

I've never seen that link before. TYVM, even aside from Lizette's original problem. :: waves at Lizette ::

-phil (sevetson)

Philip Sevetson
Computer Systems Manager
FISA-OPA
5 Manhattan West
New York, NY 10001
[login to unmask email]
212-857-1688 w
917-991-7052 m
212-857-1659 f
[cid:[login to unmask email]

From: Grainger, Phil [mailto:[login to unmask email]
Sent: Wednesday, July 10, 2019 9:09 AM
To: [login to unmask email]
Subject: [DB2-L] - RE: How to identify what is process is creating Unaccounted time in DB2

If I remember rightly, Accounting Class 3 if started gives more granularity on "unknown" wait timers (just checked the KC "Accounting class 3 data provides detailed information about the distribution of suspension times and related events.")

So I assume you are currently running without Class 3?

Also see https://www.ibm.com/support/knowledgecenter/en/SSEPEK_11.0.0/perf/src/tpc/db2z_investigatehighclass3.html

Phil Grainger
Principal Enablement Manager

[BMC Exchange 2019 - Global Event Series - REGISTER] https://www.bmc.com/ami

Direct

+44 1189 218 000

Mobile

+44 7808 643 479

Email

[login to unmask email]

E2, Eskdale Road
Winnersh
Berkshire
United Kingdom
RG41 5TS
[image001 (002)][https://acclaim-production-app.s3.amazonaws.com/images/2429c3cd-a1de-44fc-b4f3-bc762bb2f963/IBM%2BChampion%2B-%2BAnalytics%2B2018.png][https://acclaim-production-app.s3.amazonaws.com/images/2429c3cd-a1de-44fc-b4f3-bc762bb2f963/IBM%2BChampion%2B-%2BAnalytics%2B2018.png]



From: Lizette Koehler <[login to unmask email]>
Sent: 10 July 2019 13:48
To: [login to unmask email]
Subject: [EXTERNAL] [DB2-L] - How to identify what is process is creating Unaccounted time in DB2

I am trying to understand why a DB2 application runs fine when only one user is running, but gets really slow when it gets up to 50. The applications team feels they should be able to handle several hundred users at one time.

We are using DB2 PM and Detector to do the analysis

Looking for guidance on isolating and identifying what is "unaccounted time" in DB2 as identified by CA Detector

I understand it could be as simple as waiting on HSM recall. But there were no recalls when the users were executing their DB2 application.

And my very basic understanding is that DB2 is "waiting" for something to complete.

What we are seeing is a large part of the DB2 work seems to be in something called "unaccounted" time.

When we look at RMF, the CPU is not heavily used (less than 50% busy). RMF shows DB2 waiting on CPU. There is no paging going on.
No error messages from CF or HSM or other subsystems on the LPAR.

This is from a server to DB2 then back, and no messages that seem to indicate any issues within z/OS

Could someone help me understand what is "unaccounted time" is from DB2 that would be and how I would go about determining where this process is doing in DB2 when that this occurs. How can I identify what it is DB2 is requesting that is causing it to wait.

I will be doing more research on the internet on this, but thought someone could provide me with some pointers on this area.


Thank you

Lizette Koehler
statistics: A precise and logical method for stating a half-truth inaccurately


-----End Original Message-----
BMC Software Limited Registered Office: Building E2, Eskdale Road, Winnersh, Wokingham, Berkshire, United Kingdom, RG41 5TS Registered in England No. 1927903 The content of this email is confidential. If you are not the addressee, you may not distribute, copy or disclose any part of it. If you receive this message in error, please delete this from your system and notify the sender immediately.
-----End Original Message-----
**This e-mail, including any attachments, may be confidential, privileged, or otherwise legally protected. It is intended only for the addressee. If you received this e-mail in error or from someone who was not authorized to send it to you, do not disseminate, copy, or otherwise use this e-mail or its attachments. Please notify the sender immediately by reply e-mail and delete the e-mail from your system.**
Attachments

  • image002.png (3.3k)

Roy Boxwell

How to identify what is process is creating Unaccounted time in DB2
(in response to Lizette Koehler)
That is not a lot of tracing...



Roy Boxwell

SOFTWARE ENGINEERING GmbH and SEGUS Inc.
-Product Development-

Heinrichstrasse 83-85
40239 Duesseldorf/Germany
Tel. +49 (0)211 96149-675
Fax +49 (0)211 96149-32
Email: <mailto:[login to unmask email]> [login to unmask email]
Web http://www.seg.de http://www.seg.de

https://www.seg.de/corporate/rechtliche-hinweise/datenschutz Link zur
Datenschutzerklärung


Software Engineering GmbH
Amtsgericht Düsseldorf, HRB 37894
Geschäftsführung: Gerhard Schubert, Ulf Heinrich



From: Lizette Koehler [mailto:[login to unmask email]
Sent: Wednesday, July 10, 2019 4:44 PM
To: [login to unmask email]
Subject: [DB2-L] - RE: How to identify what is process is creating
Unaccounted time in DB2



Thanks.


We do have Trace 3 running (Accounting) for 01,02,03 going to SMF



Lizette







From: Bucaria, John <[login to unmask email] <mailto:[login to unmask email]> >
Sent: Wednesday, July 10, 2019 7:07 AM
To: [login to unmask email] <mailto:[login to unmask email]>
Subject: [DB2-L] - RE: How to identify what is process is creating
Unaccounted time in DB2



It has been my understanding that the Detector “UNACC_TIME” generally
referred to CPU wait but nothing more granular.

-John



From: Lizette Koehler <[login to unmask email] <mailto:[login to unmask email]> >
Sent: Wednesday, July 10, 2019 8:48 AM
To: [login to unmask email] <mailto:[login to unmask email]>
Subject: [DB2-L] - How to identify what is process is creating Unaccounted
time in DB2



I am trying to understand why a DB2 application runs fine when only one user
is running, but gets really slow when it gets up to 50. The applications
team feels they should be able to handle several hundred users at one time.



We are using DB2 PM and Detector to do the analysis



Looking for guidance on isolating and identifying what is “unaccounted time”
in DB2 as identified by CA Detector



I understand it could be as simple as waiting on HSM recall. But there were
no recalls when the users were executing their DB2 application.



And my very basic understanding is that DB2 is “waiting” for something to
complete.



What we are seeing is a large part of the DB2 work seems to be in something
called “unaccounted” time.



When we look at RMF, the CPU is not heavily used (less than 50% busy). RMF
shows DB2 waiting on CPU. There is no paging going on.

No error messages from CF or HSM or other subsystems on the LPAR.



This is from a server to DB2 then back, and no messages that seem to
indicate any issues within z/OS



Could someone help me understand what is “unaccounted time” is from DB2 that
would be and how I would go about determining where this process is doing in
DB2 when that this occurs. How can I identify what it is DB2 is requesting
that is causing it to wait.



I will be doing more research on the internet on this, but thought someone
could provide me with some pointers on this area.





Thank you



Lizette Koehler

statistics: A precise and logical method for stating a half-truth
inaccurately





-----End Original Message-----



-----End Original Message-----



-----End Original Message-----

Attachments

  • smime.p7s (5.1k)

Norbert Jenninger

RE: How to identify what is process is creating Unaccounted time in DB2
(in response to Lizette Koehler)

The „NOT ACCOUNT“ time shown in the OMEGAMON Db2 PM/PE  Accounting Report is the time which is not under control of DB2, essentially it is the difference between he Class 2 Elapsed time minus CL2 CPU minus CL3 Suspension. CL3 suspension time is the wait time of various actions under control of Db2.

You should also read the DB2 “Managing Performance” manual chapter 42 or the URL  https://www.ibm.com/support/knowledgecenter/en/SSEPEK_11.0.0/perf/src/tpc/db2z_investigatehighnotaccount.html

for more information in section “Investigating Db2 not accounted time”.  Actually you already checked some of the activities outside of DB2 described within this chapter, but you may find some additional hints there.     

 

Regards, Norbert Jenninger

Consultant at Rocket Software

Jeff Gross

RE: How to identify what is process is creating Unaccounted time in DB2
(in response to Norbert Jenninger)

The UNACCOUNTED field is not a direct measurement, but a computed field on what could not be measured.  In other words,  Total time - specific measurements = unaccounted time.  

Is this Unaccounted time the same for a single execution, a few or many?  What I am trying to understand is if this overhead linear? 

Do you see this overhead measured in the IFI fields?  If so, turn off the Detector collection and see if the behavior changes.  That is one of the recommendations in the link Norbert just provided.

 

I found a knowledge doc from the Detecotr product team and copied it below. 

//

The time incurred by the program while it is executing SQL that is not accounted for by the IFCID accounting fields. This field has no specific meaning but serves as an indication of the amount of SQL execution time not specifically accounted for by the IFCID accounting fields.  This metric is only valid when Accounting Class 3 is active.

 

This field does not correspond to any specific IFCID field. It is equivalent to IFCID field QWACASC minus the following IFCID fields:   QWACAJST, QWACAWTI, QWACAWTL, QWACAWTR, QWACAWTW, QWACAWTE, QWACAWTP, QWACAWTG, QWACAWTJ, QWACSPTT, QWACAWLG, QWACUDTT, QWACTRTT, QWACTRTE, QWACAWTK, QWACAWTM, QWACAWTN, QWACAWTO, QWACAWTQ, QWAXALOG, QWAXAWDR, QWAXAWCL, QWAXAWAR, QWAXOCSE, QWAXSLSE, QWAXDSSE, QWAXOTSE, QWAXIXLT, QWACCLS2_ZIIP, and  QWACALBW.                                          

The data Detector uses to compute this field is not directly obtained from the IBM instrumentation facility, but the field definitions are similar. Detector reports this information at different levels of granularity than the IFCID records may indicate. For more information on IFCIDs, refer to the IBM DB2 Administration Guide.    

//

 

Jeff Gross           

Phil Grainger

How to identify what is process is creating Unaccounted time in DB2
(in response to Philip Sevetson)
Isn’t Google great?

What DID we do without it?

Oh, we used to remember stuff and look the rest up in big ring binders ☹

Phil Grainger
Principal Enablement Manager

[BMC Exchange 2019 - Global Event Series - REGISTER] https://www.bmc.com/ami

Direct

+44 1189 218 000

Mobile

+44 7808 643 479

Email

[login to unmask email]

E2, Eskdale Road
Winnersh
Berkshire
United Kingdom
RG41 5TS
[image001 (002)][https://acclaim-production-app.s3.amazonaws.com/images/2429c3cd-a1de-44fc-b4f3-bc762bb2f963/IBM%2BChampion%2B-%2BAnalytics%2B2018.png][https://acclaim-production-app.s3.amazonaws.com/images/2429c3cd-a1de-44fc-b4f3-bc762bb2f963/IBM%2BChampion%2B-%2BAnalytics%2B2018.png]



From: Sevetson, Phil <[login to unmask email]>
Sent: 10 July 2019 18:16
To: '[login to unmask email]' <[login to unmask email]>
Subject: [EXTERNAL] [DB2-L] - RE: How to identify what is process is creating Unaccounted time in DB2

Phil G.,

I’ve never seen that link before. TYVM, even aside from Lizette’s original problem. :: waves at Lizette ::

-phil (sevetson)

Philip Sevetson
Computer Systems Manager
FISA-OPA
5 Manhattan West
New York, NY 10001
[login to unmask email]<mailto:[login to unmask email]>
212-857-1688 w
917-991-7052 m
212-857-1659 f
[cid:[login to unmask email]

From: Grainger, Phil [mailto:[login to unmask email]
Sent: Wednesday, July 10, 2019 9:09 AM
To: [login to unmask email]<mailto:[login to unmask email]>
Subject: [DB2-L] - RE: How to identify what is process is creating Unaccounted time in DB2

If I remember rightly, Accounting Class 3 if started gives more granularity on “unknown” wait timers (just checked the KC “Accounting class 3 data provides detailed information about the distribution of suspension times and related events.”)

So I assume you are currently running without Class 3?

Also see https://www.ibm.com/support/knowledgecenter/en/SSEPEK_11.0.0/perf/src/tpc/db2z_investigatehighclass3.html https://urldefense.proofpoint.com/v2/url?u=https-3A__www.ibm.com_support_knowledgecenter_en_SSEPEK-5F11.0.0_perf_src_tpc_db2z-5Finvestigatehighclass3.html&d=DwMFAg&c=UrUhmHsiTVT5qkaA4d_oSzcamb9hmamiCDMzBAEwC7E&r=EAGrd_qzLADPfI8dgytr8sbCG7_U9QfXwQMLgK1Zo30&m=HLaWJmPJmMB3RusgaQeVuE7KOBDlFTdKE9JYBX1h6O4&s=EeBnsPHZELJ01AkPcKsGdR3sbXAi30NOklQyEZsmt48&e=

Phil Grainger
Principal Enablement Manager

[BMC Exchange 2019 - Global Event Series - REGISTER] https://www.bmc.com/ami

Direct

+44 1189 218 000

Mobile

+44 7808 643 479

Email

[login to unmask email]<mailto:[login to unmask email]>

E2, Eskdale Road
Winnersh
Berkshire
United Kingdom
RG41 5TS
[image001 (002)][https://acclaim-production-app.s3.amazonaws.com/images/2429c3cd-a1de-44fc-b4f3-bc762bb2f963/IBM%2BChampion%2B-%2BAnalytics%2B2018.png][https://acclaim-production-app.s3.amazonaws.com/images/2429c3cd-a1de-44fc-b4f3-bc762bb2f963/IBM%2BChampion%2B-%2BAnalytics%2B2018.png]



From: Lizette Koehler <[login to unmask email]<mailto:[login to unmask email]>>
Sent: 10 July 2019 13:48
To: [login to unmask email]<mailto:[login to unmask email]>
Subject: [EXTERNAL] [DB2-L] - How to identify what is process is creating Unaccounted time in DB2

I am trying to understand why a DB2 application runs fine when only one user is running, but gets really slow when it gets up to 50. The applications team feels they should be able to handle several hundred users at one time.

We are using DB2 PM and Detector to do the analysis

Looking for guidance on isolating and identifying what is “unaccounted time” in DB2 as identified by CA Detector

I understand it could be as simple as waiting on HSM recall. But there were no recalls when the users were executing their DB2 application.

And my very basic understanding is that DB2 is “waiting” for something to complete.

What we are seeing is a large part of the DB2 work seems to be in something called “unaccounted” time.

When we look at RMF, the CPU is not heavily used (less than 50% busy). RMF shows DB2 waiting on CPU. There is no paging going on.
No error messages from CF or HSM or other subsystems on the LPAR.

This is from a server to DB2 then back, and no messages that seem to indicate any issues within z/OS

Could someone help me understand what is “unaccounted time” is from DB2 that would be and how I would go about determining where this process is doing in DB2 when that this occurs. How can I identify what it is DB2 is requesting that is causing it to wait.

I will be doing more research on the internet on this, but thought someone could provide me with some pointers on this area.


Thank you

Lizette Koehler
statistics: A precise and logical method for stating a half-truth inaccurately


-----End Original Message-----
BMC Software Limited Registered Office: Building E2, Eskdale Road, Winnersh, Wokingham, Berkshire, United Kingdom, RG41 5TS Registered in England No. 1927903 The content of this email is confidential. If you are not the addressee, you may not distribute, copy or disclose any part of it. If you receive this message in error, please delete this from your system and notify the sender immediately.
-----End Original Message-----
**This e-mail, including any attachments, may be confidential, privileged, or otherwise legally protected. It is intended only for the addressee. If you received this e-mail in error or from someone who was not authorized to send it to you, do not disseminate, copy, or otherwise use this e-mail or its attachments. Please notify the sender immediately by reply e-mail and delete the e-mail from your system.**
________________________________
Attachment Links: image002.png (3 k) https://urldefense.proofpoint.com/v2/url?u=https-3A__www.idug.org_p_fo_do_-3Fdownload-3D1-26fid-3D10277&d=DwMFAg&c=UrUhmHsiTVT5qkaA4d_oSzcamb9hmamiCDMzBAEwC7E&r=EAGrd_qzLADPfI8dgytr8sbCG7_U9QfXwQMLgK1Zo30&m=HLaWJmPJmMB3RusgaQeVuE7KOBDlFTdKE9JYBX1h6O4&s=UxEHBeiRgIM7Qw3bOHIi50reksIllmMwx7UaLQ7lwDg&e=
Site Links: View post online https://urldefense.proofpoint.com/v2/url?u=https-3A__www.idug.org_p_fo_st_-3Fpost-3D189585-26anc-3Dp189585-23p189585&d=DwMFAg&c=UrUhmHsiTVT5qkaA4d_oSzcamb9hmamiCDMzBAEwC7E&r=EAGrd_qzLADPfI8dgytr8sbCG7_U9QfXwQMLgK1Zo30&m=HLaWJmPJmMB3RusgaQeVuE7KOBDlFTdKE9JYBX1h6O4&s=QWfXP9T8WBKkjhh0of4W-YXV6tWx-oEwIiXLwfg_eYo&e= View mailing list online https://urldefense.proofpoint.com/v2/url?u=https-3A__www.idug.org_p_fo_si_-3Ftopic-3D19&d=DwMFAg&c=UrUhmHsiTVT5qkaA4d_oSzcamb9hmamiCDMzBAEwC7E&r=EAGrd_qzLADPfI8dgytr8sbCG7_U9QfXwQMLgK1Zo30&m=HLaWJmPJmMB3RusgaQeVuE7KOBDlFTdKE9JYBX1h6O4&s=aMpnU_3r8Mh_6OfwILNrwDoY1qV0DCOv0LwwQSnREMA&e= Start new thread via email<mailto:[login to unmask email]> Unsubscribe from this mailing list<mailto:[login to unmask email]?Subject=Unsubscribe> Manage your subscription https://urldefense.proofpoint.com/v2/url?u=https-3A__www.idug.org_p_us_to_&d=DwMFAg&c=UrUhmHsiTVT5qkaA4d_oSzcamb9hmamiCDMzBAEwC7E&r=EAGrd_qzLADPfI8dgytr8sbCG7_U9QfXwQMLgK1Zo30&m=HLaWJmPJmMB3RusgaQeVuE7KOBDlFTdKE9JYBX1h6O4&s=z7pyqRdP3_5gNAQG_1iendjipof433skqqUslRtcLnc&e=

This email has been sent to: [login to unmask email]<mailto:[login to unmask email]>

Discover the best cloning tool on the market. Try BCV5 & the new BCV5 Masking Tool.
ESAi also has powerful solutions for Buffer Pool Tuning, Log Analysis, TDM, & more.
http://www.ESAIGroup.com/idug https://urldefense.proofpoint.com/v2/url?u=http-3A__www.ESAIGroup.com_idug&d=DwMFAg&c=UrUhmHsiTVT5qkaA4d_oSzcamb9hmamiCDMzBAEwC7E&r=EAGrd_qzLADPfI8dgytr8sbCG7_U9QfXwQMLgK1Zo30&m=HLaWJmPJmMB3RusgaQeVuE7KOBDlFTdKE9JYBX1h6O4&s=0PkRLAfy6SEaunifN1spWmmcPJI7zNs9_KNpM6XbMmA&e=

Use of this email content is governed by the terms of service at:
http://www.idug.org/p/cm/ld/fid=2 https://urldefense.proofpoint.com/v2/url?u=http-3A__www.idug.org_p_cm_ld_fid-3D2&d=DwMFAg&c=UrUhmHsiTVT5qkaA4d_oSzcamb9hmamiCDMzBAEwC7E&r=EAGrd_qzLADPfI8dgytr8sbCG7_U9QfXwQMLgK1Zo30&m=HLaWJmPJmMB3RusgaQeVuE7KOBDlFTdKE9JYBX1h6O4&s=nSWGtoY5it_5N5BCTBKXDAo3CXVRhQuTxlJXUIjmiXY&e=

________________________________
BMC Software Limited Registered Office: Building E2, Eskdale Road, Winnersh, Wokingham, Berkshire, United Kingdom, RG41 5TS Registered in England No. 1927903 The content of this email is confidential. If you are not the addressee, you may not distribute, copy or disclose any part of it. If you receive this message in error, please delete this from your system and notify the sender immediately.
Attachments

  • image001.jpg (49.7k)
  • image002.png (6.7k)
  • image003.png (3.7k)
  • image004.png (<1k)

Avram Friedman

RE: How to identify what is process is creating Unaccounted time in DB2
(in response to Lizette Koehler)

Its multi programing didpatchrt delay and id pgyen not a valid mesurement  . Its an error on collection design

 

Hopefully you understand whu there is no multi program delay with one thing in and ready

 

Now with two things in and ready one could be us8ng and one could be waiting so we co7nt the sm3 moment of time twice

With 100 in and ready

we could have potencially have 1 Using. And 99 waiting

Boy does that look bad

Let me know if you hAve  More questions

I am the father of omegmon db2 

 

 

 

 

 

Norbert Jenninger

RE: How to identify what is process is creating Unaccounted time in DB2
(in response to Jeff Gross)

Jeff,

Regarding “Total time - specific measurements = unaccounted time.”

The Specific measurements” has to include ALLl measured wait times by DB2. If, for example, one of newer Db2 CL3 wait times is not included, e.g. QWACAAWC - ACCUMULATED WAIT TIME FOR REQUESTS TO AN ACCELERATOR.,  then this wait time is shown as part of the calculated NOT ACCOUNTed time.   

Refer to the SDSNIVPD(DSNWMSGS) - for full description of Db2 trace record IFCID 3. Db2 will measure and add new wait times IFCID counters if needed and if they could become major contributor in wait times (QWAX_PIPE_WAIT, QWAC_AT_WAIT, QWAX_LOBCOMP_WAIT, QWAX_PIPE_WAIT, …).  Those need to be considered in NOT ACCOUNT calculation depending on Db2 version/release/PTF.

Reg. "Is this Unaccounted time the same for a single execution, a few or many?  What I am trying to understand is if this overhead linear?"

The NOT ACCOUNT time can be calculated for each single thread execution (aka IFCID 3 Accounting record). So it may vary from thread to thread.

Regards, Norbert Jenninger

Consultant at Rocket Software

Avram Friedman

RE: How to identify what is process is creating Unaccounted time in DB2
(in response to Norbert Jenninger)


Let me try agan

First sorry fot my bad typing.   It is due to shaking

Assuming a uniprocrssor.,.  

When a thread changes what is doing.  All thrrads arr examed.  Please remember this is thread mesurement not db2 mesurement

 

If a thread is in and ready and 2nd on the queue it is considered to be using cpu

 ,  why 2nd ecause # one on the queue is the 

Esurement

 

Every in and rrady thread iz waiting for cpu for an unknown reason

 

A ini processor can only have one in rrady and using cpu in any instamce

 

The more i.  And teady threads at a moment the more. Wait fot unknown rrasons

 

Its about the thread) applicati9n) n9t db2

It's about queue position not CPU

Can't know about CPU usage furring a msurement cycle because the mm esurement its self is using the cpu

 

 

 

 

Avram Friedman
DB2-L hall of fame contributor
DB2-L 'past' administrator

[login to unmask email]