CAIENF SNMP trace service getaddrinfo() rc=1 after IPL
Article ID: 262472
Updated On:
Products
Issue/Introduction
The SNMP trap monitor (PGM=CAW1SNMP), the STC ENFSNMPD, failing to start after IPL (automated with SSM).
E.g.
18:26:52 MAIN: Started on Sat Mar 11, 2023 - asid(005F)
18:26:52 MAIN: Sysname: z/OS
18:26:52 MAIN: Nodename: OX02
18:26:52 MAIN: Version: 02
18:26:52 MAIN: Release: 05.00
18:26:52 mainInit: node name 'OX02' is not valid, getaddrinfo() rc=1
18:26:52 MAIN: main_init return rc=4
The solution is simply to restart it later.
Possibly you are starting it too early in the IPL sequence. We connected the startup of the STC to the EZD1214I or EZD1314I from the TCP/IP stc.
ENFSNMPD has TCPIP as prereq and now starting a little later. But to no avail. There is still this error message.
Are there other dependency? How to avoid this error from occurring?
Environment
Release : 15.0
Cause
The ENFSNMP is not dependent to the TCPIP stc to get fully initialized, but the RESOLVER stc should be fully initialized...
The RESOLVER is a separate task, besides TCP/IP and this task should be fully initialized...
Resolution
However, ENFSPNMP only works when ENF is also fully active..
Starting ENFSNMP before ENF, or before it's fully initialized is of no use...
The ENFSNMP function only works when ENF is fully active..
Advised is to make it dependent to ENF, or add it to the ENF auto command member and let ENF start it when it's initialized..
See this page :
https://techdocs.broadcom.com/us/en/ca-mainframe-software/traditional-management/ca-common-services-for-z-os/15-0/administrating/event-notification-facility-caienf/snmp-monitor.html
It mentions this startup option.. The rest of the rest of the automation can be left in tact, but let ENF arrange the startup...
Feedback
