Network Troubleshooting from the HP 3000

One of the more frequent questions we get involves network problems such as “None of my users can connect to the system.”

Our methodology for troubleshooting such a problem is to start with a :LINKCONTROL @;ALL command and look at the errors in the output. Next we execute a :LINKCONTROL @, RESET to reset the statistics. After this we know all errors in the :LINKCONTROL @;ALL output are fresh errors.

The next step in troubleshooting is to PING various pieces of the network from the HP 3000 and from a client PC. Determining which routers, switches, and workstations are answering ping requests creates a road map for finding a faulty piece of hardware.

Another powerful feature of the LINKCONTROL command is its tracing feature. A trace is started dynamically by issuing a LINKCONTROL. LANLINK;TRACE=ON command. The trace file is processed with the program NMDUMP.PUB.SYS. The trace file is comparable to a packet sniffer, but only for data destined for the HP 3000, not the entire network. The LINKCONTROL command is secured by only allowing access to users with NM capability. Since sensitive data may be transmitted as clear text, this command should be secured properly and NM capability should be handed out sparingly.

Another useful method for troubleshooting network problems is Simple Network Management Protocol (SNMP) which is a standard for managing network hardware. SNMP allows polling of status information of another network device. MPE/iX contains a powerful SNMP subsystem.