0
© 2009 IBM Corporation
zIIP Capacity Planning
Martin Packer
martin_packer@uk.ibm.com @martinpacker
© 2014 IBM Corporation2
Abstract
zIIP Capacity Planning tends to be neglected - in favour of General-Purpose Engines (GCPs...
© 2014 IBM Corporation3
If You Only Remember 3 4 Things
Capacity Planning for zIIPs has become much more important
Pay e...
© 2014 IBM Corporation4
Why People Don't Do zIIP Capacity Planning
zIIPs are much cheaper than general purpose processors...
© 2014 IBM Corporation5
But consider
Spilling to a GCP costs money
Not spilling to a GCP can cost performance & scalabil...
© 2014 IBM Corporation6
Recent News
zAAP On zIIP
– Even with zAAPs installed
New 2 : 1 zIIP : GCP configuration rule wit...
© 2014 IBM Corporation7
DB2 Version 10 Deferred Write And Prefetch Engines
Substantial portion of DBM1 address space CPU
...
© 2014 IBM Corporation8
Why Protect DB2 System Address Spaces From Being Pre-empted?
MSTR contains the DB2 system monitor...
© 2014 IBM Corporation9
© 2014 IBM Corporation10
So Let's Take This Seriously...
© 2014 IBM Corporation11
Many Sources Of Information
From RMF:
– SMF 70
– SMF 72-3
From z/OS:
– SMF 30
From middleware:...
© 2014 IBM Corporation12
With RMF data you can ... see zIIP Pool Busy
© 2014 IBM Corporation13
© 2014 IBM Corporation14
...break down by LPAR
© 2014 IBM Corporation15
© 2014 IBM Corporation16
...drive down to the Service Class level
(Slide is for zIIP on GCP)
© 2014 IBM Corporation17
© 2014 IBM Corporation18
Service Class (Period) Information
For Velocity Goal service class periods you get samples...
F...
© 2014 IBM Corporation19
© 2014 IBM Corporation20
Understanding Your Exploiters
Exploiters come in different shapes and sizes
– And leave “footpri...
© 2014 IBM Corporation21
Understanding Your Exploiters ...
Table summarised over several hours. You probably need peak 15 ...
© 2014 IBM Corporation22
A DB2 Version DBM1 Address Space's zIIP and CPU Usage Over 24 Hours
© 2014 IBM Corporation23
Projections
Projecting Near-Term zIIP CPU Requirements:
– For existing workloads that don't run ...
© 2014 IBM Corporation24
zIIP “Short Engine” Effect LPAR Configuration is important
LPARs with low zIIP usage sharing a z...
© 2014 IBM Corporation25
zIIP To GCP Ratios
What's most important is how the zIIP performs
Extreme case 12 : 1 GCP : zII...
© 2014 IBM Corporation26
Subcapacity General Purpose Processors
zIIP runs at full speed
– e.g. about twice as fast as a 6...
© 2014 IBM Corporation27
Conclusions
© 2014 IBM Corporation28
If You Only Remembered 3 4 Things I Hope They Were
Capacity Planning for zIIPs has become much m...
© 2014 IBM Corporation29
Burn Before Reading :-)
Date Version
13 Jan 2014 1 Original version
07 Feb 2014 2 Updated thanks ...
Upcoming SlideShare
Loading in...5
×

zIIP Capacity Planning

1,577

Published on

Published in: Technology
0 Comments
1 Like
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total Views
1,577
On Slideshare
0
From Embeds
0
Number of Embeds
4
Actions
Shares
0
Downloads
59
Comments
0
Likes
1
Embeds 0
No embeds

No notes for slide

Transcript of "zIIP Capacity Planning"

  1. 1. © 2009 IBM Corporation zIIP Capacity Planning Martin Packer martin_packer@uk.ibm.com @martinpacker
  2. 2. © 2014 IBM Corporation2 Abstract zIIP Capacity Planning tends to be neglected - in favour of General-Purpose Engines (GCPs). With recent enhancements to DB2 allowing you to offload critical CPU to zIIPs, and to get the most out of zAAP-on-zIIP, it's time to take zIIP Capacity Planning seriously. This presentation describes how to do zIIP Capacity Planning properly - with instrumentation and guidelines.
  3. 3. © 2014 IBM Corporation3 If You Only Remember 3 4 Things Capacity Planning for zIIPs has become much more important Pay especial attention to what % Busy represents “full” Measure zIIP potential and usage down to the address space level Consider LPAR configuration carefully for zIIPs
  4. 4. © 2014 IBM Corporation4 Why People Don't Do zIIP Capacity Planning zIIPs are much cheaper than general purpose processors (GCPs) – Maintenance cost lower also zIIPs don't attract a software bill zIIP Utilization hasn't historically been all that high zIIP-eligible work can spill to GCPs at a pinch zIIPs are sometimes not the most carefully considered part of a purchase Ignoring zIIPs, ICFs, IFLs, zAAPs simplifies things to a single processor pool
  5. 5. © 2014 IBM Corporation5 But consider Spilling to a GCP costs money Not spilling to a GCP can cost performance & scalability zIIPs aren't free Shoving another processor book in can be difficult or impossible – And at end of processor life upgrades become impossible More software that exploits zIIP installed – Some more performance-critical than previous exploiters New and modernised applications increasingly use zIIP LPAR configuration for zIIPs affects their performance
  6. 6. © 2014 IBM Corporation6 Recent News zAAP On zIIP – Even with zAAPs installed New 2 : 1 zIIP : GCP configuration rule with zEC12 and zBC12 processors DB2 Version 10 Deferred Write and Prefetch Engines DB2 Version 11 further exploiters
  7. 7. © 2014 IBM Corporation7 DB2 Version 10 Deferred Write And Prefetch Engines Substantial portion of DBM1 address space CPU – Changed by APAR PM30468 from MSTR address space 100% eligible for zIIP Very stringent performance requirements – Must not be delayed – Latency to cross over to GCP generally too high Substantially changes zIIP Capacity Planning rules Treat DBM1 as a key address space and protect its access to CPU, including zIIP
  8. 8. © 2014 IBM Corporation8 Why Protect DB2 System Address Spaces From Being Pre-empted? MSTR contains the DB2 system monitor task – Requires aggressive WLM goal to monitor CPU stalls & virtual storage constraints DBM1 manages DB2 threads & critical for local DB2 latch & cross-system locking negotiation – Any delay in negotiating a critical system or application resource (e.g. P-lock on a space map page) can lead to a slowdown of the whole DB2 data sharing group DIST & WLM-managed stored procedure AS only run the DB2 service tasks – i.e. work performed for DB2 not attributable to a single user – Classification of incoming workload, scheduling of external stored procedures, etc. – Typically means these address spaces place a minimal CPU load on the system • BUT… they do require minimal CPU delay to ensure good system wide performance – The higher CPU demands to run DDF and/or SP workloads are controlled by the WLM service class definitions for the DDF enclave workloads or the other workloads calling the SP • Clear separation between DB2 services which are long-running started tasks used to manage the environment and transaction workloads that run and use DB2 services
  9. 9. © 2014 IBM Corporation9
  10. 10. © 2014 IBM Corporation10 So Let's Take This Seriously...
  11. 11. © 2014 IBM Corporation11 Many Sources Of Information From RMF: – SMF 70 – SMF 72-3 From z/OS: – SMF 30 From middleware: – SMF 101 DB2 Accounting Trace – SMF 110 CICS Statistics Trace
  12. 12. © 2014 IBM Corporation12 With RMF data you can ... see zIIP Pool Busy
  13. 13. © 2014 IBM Corporation13
  14. 14. © 2014 IBM Corporation14 ...break down by LPAR
  15. 15. © 2014 IBM Corporation15
  16. 16. © 2014 IBM Corporation16 ...drive down to the Service Class level (Slide is for zIIP on GCP)
  17. 17. © 2014 IBM Corporation17
  18. 18. © 2014 IBM Corporation18 Service Class (Period) Information For Velocity Goal service class periods you get samples... For work that's not zIIP eligible you get e.g. Using and Delay samples for CPU and I/O – These are the sample counters that affect the velocity calculation: Velocity % = 100 x Using / (Using + Delay) With zIIP-eligible work you get better resolution – as you get Using and Delay samples specific to zIIP Useful to correlate Delay for zIIP with % zIIP on GCP – Can explain behaviour
  19. 19. © 2014 IBM Corporation19
  20. 20. © 2014 IBM Corporation20 Understanding Your Exploiters Exploiters come in different shapes and sizes – And leave “footprints” accordingly Dependent Enclave – e.g. DBM1 Prefetch / Deferred Write Independent Enclave – e.g. DDF zAAP-on-zIIP – Formerly zAAP-eligible e.g. Java e.g. Dependent Enclave for DBM1 Address Space is Performance Critical from DB2 Version 10
  21. 21. © 2014 IBM Corporation21 Understanding Your Exploiters ... Table summarised over several hours. You probably need peak 15 – 30 mins
  22. 22. © 2014 IBM Corporation22 A DB2 Version DBM1 Address Space's zIIP and CPU Usage Over 24 Hours
  23. 23. © 2014 IBM Corporation23 Projections Projecting Near-Term zIIP CPU Requirements: – For existing workloads that don't run on a zIIP: • “Project CPU” reports how much CPU is zIIP-eligible via RMF • DB2 V9 Specific: Coarse Upper Bound on V10 zIIP-eligible is all DBM1 CPU – For workloads that don't exist yet use any sizing guidelines you can get – Some “adaptive exploiters” exist: • These only try to use zIIP if one configured • “Project CPU” will show zero Projecting Into The Future – The same as standard CPU Capacity Planning
  24. 24. © 2014 IBM Corporation24 zIIP “Short Engine” Effect LPAR Configuration is important LPARs with low zIIP usage sharing a zIIP – Especially with a low weight Potential high latency in logical zIIP being dispatched – If other LPARs have the zIIP In multiple zIIP case Hiperdispatch can help – “Corrals” work into fewer logical ( & physical) zIIPs – But e.g. 1-Way Queuing regime has unhelpful characteristics Of especial concern: – Single zIIP LPAR with low weight and only “DBM1 Engines” exploiter – Consider not configuring the zIIP to the LPAR
  25. 25. © 2014 IBM Corporation25 zIIP To GCP Ratios What's most important is how the zIIP performs Extreme case 12 : 1 GCP : zIIP ratio especially unhelpful because of the “1” An unscientific view is that somewhere in the region of 2 : 1 to 4 : 1 is rarely harmful Consider both “in LPAR” and “across the machine” effects – Sometimes the answer will be “no zIIPs for this LPAR”
  26. 26. © 2014 IBM Corporation26 Subcapacity General Purpose Processors zIIP runs at full speed – e.g. about twice as fast as a 6xx General Purpose Processor (GCP) Good news: – zIIP higher capacity and faster than GCP Not so good news: – zIIP-eligible work running on a GCP is worse – Variability in speed if some runs on zIIP and some not • Not terribly different to DDF situation since PM12256 How this plays depends on numbers of threads and CPU-intensiveness CPU Conversion Factor: – Use R723NFFS to convert zIIP CPU to GCP equivalent (and vice versa) Consider configuring more zIIPs than the raw capacity requirement would suggest
  27. 27. © 2014 IBM Corporation27 Conclusions
  28. 28. © 2014 IBM Corporation28 If You Only Remembered 3 4 Things I Hope They Were Capacity Planning for zIIPs has become much more important Pay especial attention to what % Busy represents “full” Measure zIIP potential and usage down to the address space level Consider LPAR configuration carefully for zIIPs
  29. 29. © 2014 IBM Corporation29 Burn Before Reading :-) Date Version 13 Jan 2014 1 Original version 07 Feb 2014 2 Updated thanks to comments from Kathy Walsh and slide on DB2 address spaces from John Campbell. Also has abstract. 13 Apr 2014 3 Added slide on Subcapacity Models. Plus thoughts on V10 Estimation and on % of DBM1 work that is zIIP-eligible.
  1. A particular slide catching your eye?

    Clipping is a handy way to collect important slides you want to go back to later.

×