6. DATA SOURCES › 6.8 PR/SM LPAR Concepts › 6.8.5 Data Analysis › 6.8.5.1 PR/SM Related Hardware and SCP Analyzer Files › 6.8.5.1.2 HARVPA/HARLPC Relationship › 6.8.5.1.2.1 zSeries and Earlier Models' Data Organization
6.8.5.1.2.1 zSeries and Earlier Models' Data Organization
On zSeries and earlier models, PR/SM manages processor
resources in two pools: the standard CP processors and the
special purpose processors. It is not able to distinguish
between the various physical engines' types within the
special purpose processors' pool (Internal Coupling
Facilities, Integrated Facilities for Linux, or zSeries
Application Assist Processors). For PR/SM, they all belong
to the generic ICF category. Fortunately, although not
available in the SMF data, this information can be obtained
by examining the configuration of a logical partition, and
included in the HARVPA file's key sequence, providing an
additional insight into the generic special purpose
processors' pool.
There is no parallel way, for the HARLPC file, to accurately
provide more detailed information about a logical partition
purpose, like its system control program for example, than
its type (general purpose or special purpose LPAR). This is
the reason why the structure and organization of this file
sticks to the processor pool approach, providing separate
sets of measurements for standard CP processors and special
purpose (ICF) processors.
Note that the HARLPC file contains a separate specialized
processor management flag, LPCSEPPU, that can be used to
determine the scope of ICF processors, related data elements:
For zSeries and earlier models, LPCSEPPU is equal to N,
indicating that the data elements related to ICF processors
encompass the activity of all specialized processors
(internal coupling facilities, IFLs, and zAAPs).
Let's have a closer look at the HARVPA/HARLPC files'
organization. If we speak in terms of tables with rows and
columns, we can say that the HARVPA file contains one row per
logical processor configured within a logical partition. In
each row, one column is used to identify whether the
remaining metrics pertain to a standard CP or special purpose
(ICF) processor, and another column further identifies the
type of physical engine dispatched to this special purpose
processor (ICF, IFL, or IFA for zAAP). The same information
is represented in the HARLPC file by a single row for each
logical partition with two sets of summarized data columns:
one for standard CP processors' metrics, and one for special
purpose (ICF) processors, regardless of their physical types.
Example
-------
Let's assume the following PR/SM configuration:
o ZOS1 is z/OS LPAR configured with two standard CP
processors and one zAAP.
o ZOS2 is z/OS LPAR configured with two standard CP
processors.
o CF1 is a coupling facility LPAR configured with one
Internal Coupling Facility (ICF).
o LNX1 is Linux only LPAR configured with one Integrated
Facility for Linux (IFL).
HARVPA structure
----------------
LPAR Log. Proc Log Proc Engine Processor
Name Address Type Type Busy Time
---- --------- -------- ------ ---------
ZOS1 0 CP CP 07:00
ZOS1 1 CP CP 07:00
ZOS1 2 ICF IFA 12:00
ZOS2 0 CP CP 03:00
ZOS2 1 CP CP 03:00
CF1 0 ICF ICF 15:00
LNX1 0 ICF IFL 15:00
HARLPC structure
----------------
LPAR LPAR Std CP ICF Procs
Name Type Busy Time Busy Time
---- ---- --------- ---------
ZOS1 CP 14:00 12:00
ZOS2 CP 06:00 00:00
CF1 ICF 00:00 15:00
LNX1 ICF 00:00 15:00
For ZOS1, busy times for processors 0 and 1 have been
consolidated in the overall busy time for all standard CP
processors (same for ZOS2); and busy time for the zAAP
(processor 2) has been summarized in the ZOS1 overall busy
time for all processors belonging to the special purpose
(ICF) processors' pool. CF1 with its Integrated Coupling
Facility only contributes to the special purpose pool, as
LNX1 does with its Integrated Facility for Linux.
On zSeries and earlier models, PR/SM manages processor
resources in two pools: the standard CP processors and the
special purpose processors. It is not able to distinguish
between the various physical engines' types within the
special purpose processors' pool (Internal Coupling
Facilities, Integrated Facilities for Linux, or zSeries
Application Assist Processors). For PR/SM, they all belong
to the generic ICF category. Fortunately, although not
available in the SMF data, this information can be obtained
by examining the configuration of a logical partition, and
included in the HARVPA file's key sequence, providing an
additional insight into the generic special purpose
processors' pool.
There is no parallel way, for the HARLPC file, to accurately
provide more detailed information about a logical partition
purpose, like its system control program for example, than
its type (general purpose or special purpose LPAR). This is
the reason why the structure and organization of this file
sticks to the processor pool approach, providing separate
sets of measurements for standard CP processors and special
purpose (ICF) processors.
Note that the HARLPC file contains a separate specialized
processor management flag, LPCSEPPU, that can be used to
determine the scope of ICF processors, related data elements:
For zSeries and earlier models, LPCSEPPU is equal to N,
indicating that the data elements related to ICF processors
encompass the activity of all specialized processors
(internal coupling facilities, IFLs, and zAAPs).
Let's have a closer look at the HARVPA/HARLPC files'
organization. If we speak in terms of tables with rows and
columns, we can say that the HARVPA file contains one row per
logical processor configured within a logical partition. In
each row, one column is used to identify whether the
remaining metrics pertain to a standard CP or special purpose
(ICF) processor, and another column further identifies the
type of physical engine dispatched to this special purpose
processor (ICF, IFL, or IFA for zAAP). The same information
is represented in the HARLPC file by a single row for each
logical partition with two sets of summarized data columns:
one for standard CP processors' metrics, and one for special
purpose (ICF) processors, regardless of their physical types.
Example
-------
Let's assume the following PR/SM configuration:
o ZOS1 is z/OS LPAR configured with two standard CP
processors and one zAAP.
o ZOS2 is z/OS LPAR configured with two standard CP
processors.
o CF1 is a coupling facility LPAR configured with one
Internal Coupling Facility (ICF).
o LNX1 is Linux only LPAR configured with one Integrated
Facility for Linux (IFL).
HARVPA structure
----------------
LPAR Log. Proc Log Proc Engine Processor
Name Address Type Type Busy Time
---- --------- -------- ------ ---------
ZOS1 0 CP CP 07:00
ZOS1 1 CP CP 07:00
ZOS1 2 ICF IFA 12:00
ZOS2 0 CP CP 03:00
ZOS2 1 CP CP 03:00
CF1 0 ICF ICF 15:00
LNX1 0 ICF IFL 15:00
HARLPC structure
----------------
LPAR LPAR Std CP ICF Procs
Name Type Busy Time Busy Time
---- ---- --------- ---------
ZOS1 CP 14:00 12:00
ZOS2 CP 06:00 00:00
CF1 ICF 00:00 15:00
LNX1 ICF 00:00 15:00
For ZOS1, busy times for processors 0 and 1 have been
consolidated in the overall busy time for all standard CP
processors (same for ZOS2); and busy time for the zAAP
(processor 2) has been summarized in the ZOS1 overall busy
time for all processors belonging to the special purpose
(ICF) processors' pool. CF1 with its Integrated Coupling
Facility only contributes to the special purpose pool, as
LNX1 does with its Integrated Facility for Linux.
