Analysis
- Are you connected to the correct TCP/IP address space? This is
obviously a concern when running multiple stacks. See Problems connecting SNMP agents to multiple TCP/IP stacks.
If you get a message “unable to connect to TCPIP JOBNAME,” you are not connected to the correct address space. If you have defined two or more stacks, make sure your TCPIPjobname in the TCPIP.DATA data set used by the SNMP agent matches the NAME field on the SUBFILESYSTYPE statement for ENTRYPOINT(EZBPFIN) in the BPZPRMxx member you used to start z/OS UNIX MVS™.
- Did any socket-related errors occur?
Check the SNMP agent syslogd for socket(), bind(), accept(), or other socket error messages. For example, a bind() failure occurs when one or more of the ports needed by the SNMP agent is already in use. See z/OS Communications Server: IP Configuration Guide for more information about syslogd.
- Is the correct TCPIP.DATA information being used? Is the SYSTCPD DD statement coded in the PROC JCL? Is the RESOLVER_CONFIG environment variable passed on the SNMP agent initialization parameters?
- Dump of SNMP agent address space.
- Dump of TCP/IP address space.
- The syslogd traces from the agent (using trace level 255). See z/OS Communications Server: IP Configuration Guide for more information about reading the syslogd.
Information about obtaining a dump can be found in z/OS MVS Diagnosis: Tools and Service Aids for your release of MVS. Obtaining SNMP traces is discussed in SNMP traces.