Determining the network problem

Because businesses depend so heavily upon the availability of data processing systems, a problem in the network can be catastrophic. Real money is lost when networks fail. When the network is used for connecting a transaction processing mainframe to the outside world, the losses can be staggering. Keeping the network problem-free and responsive is a priority, so determining where a problem lies and fixing it quickly is imperative.

Your first indication of a network problem may come from users or operators. Users might complain about:

An unusual message being received
System behavior being different than the past
Slow response time on the network
No response from the system

System operators might report:

Poor performance
A TCP/IP or VTAM abnormal end (abend)
A loop or wait on the subsystem (TCP/IP or VTAM)

Do the following:

Read messages in the system log. Each z/OS component has a unique message prefix that identifies the component.
- VTAM messages are prefixed with IST.
- TCP/IP messages are prefixed with EZ.
- CSM messages are prefixed with IVT.
- TSO messages are prefixed by IKT.
Look for suffix E, which stands for error; for example:
```
IST999E VTAM MESSAGE LOST - INSUFFICIENT STORAGE
```
Check the appropriate messages manual for an explanation of the error.
Note: Unsure of the appropriate messages manual? Try the following web site:
http://www.ibm.com/servers/eserver/zseries/zos/bkserv/lookat/
Also check the system log to see if the system created an internal memory dump at error time.
Find out whether the system has changed; for example, whether the procedures for any components have changed.
- VTAM procedure
- VTAM configuration (VTAMLST)
- TCP/IP procedure
- TCP/IP profile data set
Check whether there has been a recent hardware change to:
- The central processor complex (CPC)
- Channel paths
- OSA-Express (Open System Adapter Express) type or definition
- Network equipment

After you locate the general problem area, use the tools and diagnostic aids at your disposal to track down the problem.