How to diagnose SOLVE/NetMaster using High CPU on Mainframe ?
Release:
Component:
Diagnosing High CPU conditions:
- does the CPU usage spike or remain constant ?
- if spiking does it occur on a regular interval basis ?
- Did the condition occur after invoking a specific screen in Netmaster ?
- How long has the region been up ?
- Is the high CPU condition causing other issues on your system ?
To run CPU monitoring on SOLVE/NetMaster region (called "Minute Minder"),
Please follow the following instructions:
XOPT=(SDUMP,CPUMM,CPUPM,CPULM)
This will report all CPU activity in the region.
Also add this:
PPREF='XM=BEST'
to your RUNSYSIN parameters. This will use the zIIP if that is the best mode.
$$SYSPRO PMON 120
SHOW NCL=ALL
SHOW NCLSTAT
This command will take a report every 120 minutes. If you do not specify 120 it will default to 10 minutes.
EV 00.30.00 LIMIT=48 KEEP=SYS ROUTE=SYS CMD=START xxxxxxx
(NCL procedure containing the $$SYSPRO PMON and SHOW commands).
This will provide a sampling on CPU usage every half of hour and running over 24 hours.
The result is in the activity log. Please return a copy when you get this the day after.
Please terse and FTP the data to Broadcom.
Note that now when you open a new case the ftp directories are automatically created so you can upload data straight away.
Please terse (IBM utility TRSMAIN) these files before FTPing to make upload easier.
Please remember to FTP in Binary!