Technical data

Configuring SNMP
13.6 Solving SNMP Problems
6. Check the log files for any problems associated with SNMP startup. For
detailed information, start the SNMP components separately with tracing
enabled, as described in Section 13.6.4.
7. Use a protocol analyzer to intercept messages going to the target. The
TCPTRACE utility is available on OpenVMS hosts. Enter the DCL command
HELP TCPTRACE for information about how to use this utility. For the
failing message:
Confirm the community configuration, as described in Section 13.6.5.2.2.
Make sure the default community is configured correctly. For example,
make sure that a read-only community name, such as ‘‘public,’ is not
being used for
set
requests. For more information about community
names, refer to the Compaq TCP/IP Services for OpenVMS SNMP
Programming and Reference guide.
Check to make sure the client used the correct query format.
8. Check for problems with ownership, protections, or installation of images,
using standard OpenVMS DCL commands, such as DIRECTORY and
INSTALL.
For example, the following message indicates that one of these factors is a
possible problem:
WARNING: select returned -1 on snmpd sockets: not owner
The owner for all SNMP executables should be [SYSTEM]. At a minimum,
the protection should be set to S:RWED,O:RWED,G:RE,W:RE.
9. If you cannot get a response for MIB variables handled by certain subagents,
verify that the subagents are running by entering the following command:
$ SHOW SYSTEM
Check for the following processes:
TCPIP$SNMP_n (master agent)
TCPIP$OS_MIBS (standard subagent)
TCPIP$HR_MIB (standard subagent)
See the Compaq TCP/IP Services for OpenVMS SNMP Programming and
Reference guide for descriptions of these processes.
Also check for custom subagents whose process names appear after RUN
commands in the following command procedure:
SYS$SYSDEVICE:[TCPIP$SNMP]TCPIP$EXTENSION_MIB_RUN.COM.
If these processes and additional subagents follow the model of the Chess
example, they should be in LEF state. Excessive time in HIB state
indicates a problem. If the processes are not there, check log files for
the possible cause of abnormal termination. Note that you must run the
SYS$STARTUP:TCPIP$SNMP_SHUTDOWN.COM procedure in order to
see entries in the latest .LOG and .ERR files. If a query on members of the
hrFSTable
group results in no response or in a ‘‘no such name’’ response, the
problem might be one of the following:
No devices are mounted through NFS.
Access to mount information is not available because the proxy is not set
up to allow the TCPIP$HR_MIB subagent to access NFS-mounted disks.
13–28 Configuring SNMP