coupling facility cpu
DESCRIPTION
Short presentation I gave to the UKCMG 1-day mini-conference 15 October in London. Covers 2 main aspects of Parallel Sysplex performance, both in the CPU area: 1) Comparing Type 70 view of CPU to Type 74-4. 2) Type 74-4 Structure-Level CPU and its role in Capacity Planning and Performance.TRANSCRIPT
© 2009 IBM Corporation 1
What's New With Coupling Facility CPU
Martin Packer
[email protected]+447802245584Twitter: MartinPacker
© 2009 IBM Corporation2
Abstract
Recently RMF's reporting of Coupling Facility CPU was enhanced, mainly to give more granularity.
This presentation outlines the author's experience with this important new instrumentation, both from the perspective of Capacity Planning and from the perspective of how parallel sysplexes perform under increasing load. It also covers other areas of Parallel Sysplex performance.
IN CASE YOU WERE IN ANY DOUBT: “other areas” does not mean “ALL other areas”. :-)
© 2009 IBM Corporation3
Topics
Structure-Level CPU
– Structure CPU Experiment
CPU / LPAR Match Up Between 70-1 and 74-4
Conclusings and Musions
© 2009 IBM Corporation 4
Structure-Level CPU
© 2009 IBM Corporation5
Structure-Level CPU Consumption CFLEVEL 15 and z/OS R.9
– Most customers are now this far advanced
New SMF 74-4 Field: R744SETM
– “Structure Execution Time”
Always 100% Capture Ratio
– Adds up to R744PBSY
Multiple uses:
– Capacity planning for changing request rates
– Examine which structures are large consumers
– Compute CPU cost of a request
• And compare to service time• Interesting number is “non-CPU” element of service time - as we shall see
– Understand whether CPU per request has degraded
– Estimating Structure Duplexing cost
NOTE:
– Need to collect 74-4 data from all z/OS systems sharing to get total request rate
• Otherwise “CPU per request” calculation will overestimate
© 2009 IBM Corporation6
Structure CPU Experiment
© 2009 IBM Corporation7
Structure CPU Experiment Based on
– R744SETM Structure Execution Time– Sync Request Rate
• Virtually no Async– Sync Service Time
One minute RMF intervals– Sorted by request rate increasing
Run was 1-way DB2 Datasharing– Only really active structures ISGLOCK and LOCK1
Red lines are CPU time per request– Blue lines are Service time per request
ISGLOCK “low volume”– Shows amortization of some fixed cost effect
• Wondering also if some “practice effect” affects service times– CF used IC links
LOCK1 “high volume”– More reliable for capacity planning– CF used a mixture of ISC and ICB links
© 2009 IBM Corporation8
ISGLOCK Requests
0
2
4
6
8
10
12
14
16
0 10 20 30 40 50 60 70
Requests / Second
Mic
rose
cond
s
CPU Time Service Time
3us?
© 2009 IBM Corporation9
LOCK1 Requests
0
2
4
6
8
10
12
750 800 850 900
Requests / Second
Mic
rose
cond
s
CPU Time Service Time
3.5us?
© 2009 IBM Corporation10
And From My Travels... Next chart isn't from the experiment just described
– A real customer system
A Group Buffer Pool
ISC-Connected
– Necessary for the customer's estate
Clearly something goes wrong at about 1100 requests / second
– Especially in response time terms but also CPU
• (Coupling Facility not CPU constrained)
Options include
– Managing the request rate to below 1100 / sec
– Working on the request mix
– Infrastructure reconfiguration
© 2009 IBM Corporation11
25us?
© 2009 IBM Corporation12
CPU / LPAR Match Up Between 701 and 744
© 2009 IBM Corporation13
●Managed out of Pool 5 in z9 and z10● Pool numbers given in SMF 70 as index into table of labels● Recommendation: Manage in reporting as a separate pool
●Follow special CF sizing guidelines● Especially for takeover situations
●Always runs at full speed● So good technology match for coupled z/OS images on same footprint● Another good reason to use ICFs is IC links
●Shared ICFs strongly discouraged for Production● Especially if the CF image has Dynamic Dispatch turned on
●Should not run ANY coupling facility above 50% busy● Especially if we need to be able to recover structures onto it
Internal Coupling Facility - Basics
© 2009 IBM Corporation14
ICF CPU Instrumentation
SMF 744 view different from SMF 701 LPAR view of processor busy•R744PBSY is CPU time processing requests•R744PWAI is CPU time while CFCC is not processing requests but it is still using CF cycles
•For Dynamic Dispatch PWAI is time when not processing CF requests but Logical CP not yet taken back by PR/SM
•For dedicated or nonDynamic Dispatch cases sum is constant•For Dynamic Dispatch sum can vary.
Number of defined processors is number of CF Processor Data sections in 744•Refined for CFLEVEL 15 by new fields for dedicated (R744FPDN) and shared (R744FPSN) processors•Also whether individual engine is dedicated (R744PTYP) and its weight (R744PWGT)
PBSY and PWAI Can be examined down to Coupling Facility engine levelSMF 744 has much more besides CF CPU instrumentation
© 2009 IBM Corporation15
CF LPAR Identification In SMF 70-1 Is Complex
Need to match LPARs in SMF 70-1 with coupling facilities in SMF 74-4 to get proper CPU picture
Since z/OS Release 8 74-4 has machine serial number
– Allows correlation in most cases
– But LPAR names and CF Names often don't match
– Often multiple CF's in same footprint with similar configuration
– Sometimes there are multiple CF's with the same name
– My code – in extremis – uses the presence of IC links to determine “colocality”
– I'm slowly learning :-) not all CF LPARs are in Pool 5
© 2009 IBM Corporation16
New Instrumentation - OA21140
Introduced to support zHPF
– Has other SMF and reporting improvements
• HiperDispatch Vertical Polarisation indicators at ENGINE level – Type 70
• Normalisation factor for zIIP – Type 70
Adds CF LPAR Partition Number
– Allows matching with SMF 70-1
RMF Level (SMFxxSRL) changed to X'55'
© 2009 IBM Corporation17
Conclusings and Musions
I think we've come a long way with Coupling Facility CPU
– Capacity Planning is now down to the structure level
• But not to the structure-by-system level
– We can now tie up the Coupling Facility and LPAR views of CPU
• With a few “corner cases”
I'd encourage you to revisit your Parallel Sysplex reporting
– Including for all the other aspects we didn't have time for
Shouldn't machines be self-documenting in SMF?