

Using the Realtime Monitor › Program Pool Detail (PF10)
Program Pool Detail (PF10)
PM-Rnn.n SYSTEM71 CA, Inc. V71 08.348 14:53:43.90
CMD--> Window : 02
Refresh: 10
02 Program Pool Detail >
Pool Total Storage High Ovrlay Ovrlay Novrlay Times Pages
Type Storage In Use Water In Use Not Used Loads Waited Loaded
PROG POOL 53248 20480 53248 3 21
REENT POOL 2456kB 488kB 488kB 42 976
XA PROG POOL 200kB 0 0
XA REENT POOL 2640kB 1592kB 1592kB 112 3184
PM-Rnn.n SYSTEM71 CA, Inc. V71 08.348 14:53:43.90
CMD--> Window : 02
Refresh: 10
02 Program Pool Detail <
Pool Storage High Ovrlay Ovrlay Novrlay Times Pages Load
Type In Use Water In Use Not Used Loads Waited Loaded Count
PROG POOL 20480 53248 3 21 3
REENT POOL 488kB 488kB 42 976 42
XA PROG POOL 0 0
XA REENT POOL 1592kB 1592kB 112 3184 112
Screen Description
The Program Pool Detail screen displays a line of information for each type of program pool defined to the system:
- 24-bit, nonreentrant program pools (PROG POOL)
- 24-bit, reentrant program pools (REENT POOL)
- 31-bit, nonreentrant program pools (XA PROG POOL)
- 31-bit, reentrant program pools (XA REENT POOL)
If your system has no reentrant pool, the DC/UCF system assigns reentrant programs to the program pool.
What To Look For
- Look at the number of times the system had to overlay active programs (Ovrlay In Use) and the number of waits (Times Waited). Any number in the Ovrlay In Use field or a large number in the Times Waited field indicates a problem with the size of the program pool. Try the options listed below to alleviate the problem:
- Increase the size of the pool. If you must increase the size of one pool, try decreasing the size of a pool that is not experiencing waits. If space is too tight to increase the size of the pool, try reducing the value assigned to the MAX TASKS and MAX ERUS parameters. This reduces concurrent demand on the program pool.
- Define heavily used, reentrant, or quasireentrant programs as resident.
- Decrease the size of frequently used programs by creating application-specific subschemas to avoid using large, global subschemas, or by segregating logical threads in applications by module.
- Compare the number of pages loaded (Pages Loaded) to the total number of loads (Load Count). If there are many pages loaded for only a few program loads, consider increasing the block size for the program load libraries and load areas. For example, a block size of 10Kb requires 20 I/Os to load a 200Kb program; a block size of 1Kb requires 200 I/Os to load the same program.
Copyright © 2014 CA.
All rights reserved.
 
|
|