SNMPD starts without errors but is not listening on any port?

Question

HP ProLiant DL360 G7 running CentOS6 5.7 with Net-SNMP 5.3.2.2-14.

I have 7 of these servers in a remote site all running mixture of CentOS 5.7 and RHEL 5.7 with same Net-SNMP installed from RPM. On one server my NMS continuously reports that it has failed to connect to the SNMP service. I ignored it because it usually clears up within ~15 minutes but today it has been happening 2-3 times per hour since the morning.

I have modified the snmpd startup script to log to file (default is to /dev/null). There are no errors in the log file upon startup:

could not open /proc/net/if_inet6
cannot open /proc/net/snmp6 ...
cmaX: listening for subagents on port 25375
cmaX: sent ColdStarts on ports 25376 to 25393
cmaX: subMIB 1 handler has disconnected
NET-SNMP version 5.3.2.2
cmaX: subMIB 2 handler has disconnected
Connection from UDP: [127.0.0.1]:50654
cmaX: subMIB 3 handler has disconnected
Received SNMP packet(s) from UDP: [127.0.0.1]:50654
cmaX: subMIB 5 handler has disconnected

I presume if_inet6 is for IPv6 which we do not use. snmpd.conf is below:

# Following entries were added by HP Insight Management Agents at
#      Tue May 15 10:58:17 CLT 2012
dlmod cmaX /usr/lib64/libcmaX64.so
rwcommunity public 127.0.0.1
rocommunity public 127.0.0.1
rwcommunity 3adRabRu 172.16.37.37
rocommunity 3adRabRu 172.16.37.37
trapcommunity traps
trapsink 172.16.37.37 traps
syscontact Lukasz
syslocation Santiago, Chile
# ---------------------- END --------------------
com2sec rwlocal default public
com2sec rolocal default public
com2sec subnet  default 3adRabRu
group   rwv2c   v2c             rwlocal
group   rov2c   v2c             rolocal
group   rov2c   v2c             subnet
view    all     included        .1
access  rwv2c   ""      any             noauth          exact   all     all     none
access  rov2c   ""      any             noauth          exact   all     none    none

Any suggestions? IP Tables is shutoff.

Solution

Check out the man page for snmpd.conf -

  agentaddress [<transport-specifier>:]<transport-address>[,...]
          defines  a  list  of listening addresses, on which to receive incoming 
          SNMP requests.  See the section LISTENING ADDRESSES in the
          snmpd(8) manual page for more information about the format of listening address

So try adding something like this to the top of snmpd.conf:

Listen for connections on all interfaces (both IPv4 and IPv6)

agentAddress udp:161,udp6:[::1]:161

Comments

[rnxrx]

Anything possibly different around SELinux setup, the way you installed snmpd or a possible sub-version? It wouldn't surprise me if they had changed the default behavior of the config file. The configuration file calls out the default as being listen on all ports but perhaps they decided that was potentially dangerous.