CA MICS Hardware and SCP Analyzer (RMF) Queries › Mainframe CPC Engine Utilization › RMFEHH – CPC Shared Engine Pools Percent Dispatch High Hours

RMFEHH – CPC Shared Engine Pools Percent Dispatch High Hours

Overview

The RMFEHH query shows the peak hour of dispatch, over the past 32 days, for each engine type (CP, zIIP, zAAP, IFL, ICF) in the shared LPAR engine pools. The query generates two dispatch charts for each engine type in each Central Processing Complex (CPC).

The example shows the two charts generated for the shared CP engine pool of an IBM z/10 2097-712 CPC. The top chart shows the peak hour, per day percent utilization of the CP engine pool, with the breakdown of the various LPARs that use CP engines. The second chart shows the same information, minus the LPAR breakdown. The chart includes a reference line (right Y-axis) that shows the number of engines in the CP engine pool (12 engines in this example). The X-axis shows the date concatenated to the high use hour. In the example, the first vertical bar shows that the peak CP engine dispatch occurred on August 29th, 2010 (a Sunday) during hour 23: "2010-08-29-23"—when the utilization of the CP engine pool was about 43 percent.

A similar set of charts is created for each engine type, one where the vertical bars show individual LPAR percent utilization, and the other showing percent utilization plus the physical engine count for the engine type.

Uses

This query is used to examine the peak utilization load for the CPC shared engine pools over the past month. Consistent peak hour utilization at, or near 100 percent means that, for at least 1 hour each day an engine pool has reached the utilization limit. It is likely that some workloads are being delayed, or receiving suboptimal attention. Consistent peak hour utilization significantly below 100 percent utilization means that you have invested more than necessary in a particular engine type.

Data Extracts

CPC Shared CP Pool Pct Use by LPAR HH

Left Y-axis

LPCPCSSU - LPAR Shared CP Processors Utilization

A user step analyzes HARLPC file data to determine the specific hour for each day where the Central Processing Complexes CP engines in the shared engine pool were most heavily dispatched. Once determined, the percent utilization of each LPAR using the CP engine pool during that peak hour is stacked in a vertical bar using data element LPCPCSSU.

X-axis

CP_DDHH - CP Engine High Use Hour for DATE

Data element CP_DDHH is derived by appending the high use hour to the date, in the form:

YYYY-MM-DD-HH (for example, 2010-08-31:22, meaning hour 22 of August 31st, 2010).

CPC Shared CP Pool Pct Use High Hour

Left Y-axis

LPCPCSSU - LPAR Shared CP Processors Utilization

The same overall shared CP engine pool utilization as shown in the previous data extract, without the individual LPAR breakdown. LPCPCSSU is summarized to the CPC level.

Right Y-axis

PRSMACSP - Avg CP Shareable Processors

Data element PRSMACSP shows the average number of CP engines in the shared pool for the high use hour. If the number of physical engines in the shared CP pool changes, this value changes over the days plotted.

X-axis

CP_DDHH - CP Engine High Use Hour for DATE

Data element CP_DDHH is derived by appending the high use hour to the date, in the form:

YYYY-MM-DD-HH (for example, 2010-08-31:22, meaning hour 22 of August 31st, 2010).

Note: This query generates eight additional data extracts, two for each of the four other engine types, ICFs, IFLs, zAAPs, and zIIPs. The charts, and methods used to generate them are identical to those used to build the CP engine pool charts. The remaining data extract documentation lists the specific data elements used for the left and right Y-axis:

• CPC Shared ICF Pool Pct Use by LPAR HH

  Left Y axis

  LPCPCSIU - LPAR Shared ICF Processors Utilization

• CPC Shared ICF Pool Pct Use High Hour

  Left Y axis

  LPCPCSIU - LPAR Shared ICF Processors Utilization

  Right Y axis

  PRSMAISP – Avg ICF Sharable Processors

• CPC Shared IFL Pool Pct Use by LPAR HH

  Left Y axis

  LPCPCSNU - LPAR Shared ICF Processors Utilization

• CPC Shared IFL Pool Pct Use High Hour

  Left Y axis

  LPCPCSNU - LPAR Shared ICF Processors Utilization

  Right Y axis

  PRSMANSP – Avg ICF Sharable Processors

• CPC Shared zAAP Pool Pct Use by LPAR HH

  Left Y axis

  LPCPCSZU - LPAR Shared zAAP Processors Utilization

• CPC Shared zAAP Pool Pct Use High Hour

  Left Y axis

  LPCPCSZU - LPAR Shared zAAP Processors Utilization

  Right Y axis

  PRSMAZSP – Avg zAAP Sharable Processors

• CPC Shared zIIP Pool Pct Use by LPAR HH

  Left Y axis

  LPCPCSDU - LPAR Shared zIIP Processors Utilization

• CPC Shared zIIP ICF Pool Pct Use High Hour

  Left Y-axis

  LPCPCSDU - LPAR Shared zIIP Processors Utilization

  Right Y-axis

  PRSMASSP – Avg zIIP Sharable Processors

Query Defaults

• HARLPC DAYS timespan file: Cycles 01 - 32
• Filters: LPCPCSxU GT 0 selected for each engine type
• User code excludes today’s data (IF DATEPART(ENDTS) EQ TODAY()-1 ;
• Derivations:

  xx_DDHH

  xx_DDHH = PUT(YEAR4,Z4.) || '-' || PUT(MONTH,Z2.) || '-' ||
  PUT(DAY,Z2.) || '-' || PUT(MAXxxH,Z2.) ;DATEPART(ENDTS)
  

  xx Engine High Hour for DATE element derived for each engine type (xx)

  CPCID

  TRIM(LPCMOD)||'-'||TRIM(CPCMODID)||'-Serial: '||SUBSTR(CPCSEQNB,12,5);
  

Modifications

The following modifications can be made to the RMFEHH query:

• You can change the HARLPC DAYS timespan cycles—defaults to cycles 1-32.
• You can use a filter statement to restrict data to focus on a particular time range (for example, HOUR GE 8 AND HOUR LE 17) or particular time zone (for example, ZONE EQ 1).
• You can add filter statements to select only a single CPC, or set of CPCs using the identifying data elements shown in the CPCID derivation.
• You can delete data extract steps for engine types not found in your CPCs. For example, if you have no IFL engines, you can delete the two data extract steps for IFL engines.