SNMP agent is not Responding

[protocult]

Hello All,

Need some help to figure out an issue. We use NIMSOFT server as a traphost to send out alerts for our NetApp environment.

Often we keep recieving these alerts "The SNMP Agent '10.zz.yy.xx' is not responding. Profile [10.zz.yy.xx]" The ip address in the message is the cluster-mgmt lif address.

Our SNMP is set up correctly, the cluster-mgmt lif, traphost server all seems to be ok. Is there any way I can zero down on the issue? We keep getting this once a day for each cluster.

Details

mt-everest::> snmp show
contact:

location:
HIMALAYAS
authtrap:
0
init:
1
traphosts:
mt-everest.himalayas.com (mt-everest.himalayas.com ) <10.xx.yy.zzz>
community:
ro public

mt-everest::> system services firewall policy show -service snmp
Policy Service Action IP-List
---------------- ---------- ------ --------------------
cluster
snmp allow 0.0.0.0/0
data
snmp deny 0.0.0.0/0
intercluster
snmp deny 0.0.0.0/0
mgmt
snmp allow 0.0.0.0/0
4 entries were displayed.

Cheers

[YIshikawa]

In general,the message "SNMP Agent ... Is not responding" means your SNMP manager software is trying to poll SNMP data from cluster management lif.

You have configured the SNMP manager not only for receiving traps  but also for SNMP polling. Please check your SNMP manager and disable SNMP polling.

[protocult]

Thanks for the information YIshikawa.  How do I check if polling is enabled or disabled. 

When I look at "event snmphistory" I do see a lot of  "app.log.info" being sent  frequently to the traphost. Does this mean that polling is enabled ?

[YIshikawa]

It is matter of SNMP manager, so ask your SNMP manager admin to check if SNMP polling for your cluster is configured on SNMP manager.

"events snmphistory show" displays  a list of SNMP trap history records, so output of this command has no relation to SNMP polling.

[protocult]

Thank you. I'll speak to them and revert.

[protocult]

If the SNMP manager is trying to poll, why is it failing? I do see that my clust_mgmt lif can talk to the SNMP management station. Is there any reason why this wont work.

[YIshikawa]

firewall policy seems OK, so any other cause like resticted network or SNMP community mismatch may prevent SNMP poling access.
Capturing packets on both SNMP manager host and cluster management LIF node may help to isolete the problem.

[protocult]

thank you. I'll post back when I hear from our nimsoft folks.