SNMP Issues

[clayman2]

I am hoping someone can help me out on this one, I have read through the forums and google trying to figure this out. I have 3 servers that I get SNMP error. Two of these servers are in a cluster, but I do not think that would matter.

The first server in the cluster, says SNMP error but shows that it is up, from the cacti box if I run snmpwalk -v 1 -c COMMUNITY_NAME IP_ADDRESS, it returns nothing, I can ping the server and anything else. I have triple checked my SNMP settings and they are correct.

The second server in the cluster says SNMP erorr but shows that it is down. If I run the same snmpwalk command on that server I get End of MIB

The third server which is a stand alone, I get SNMP error and shows a down state. If I run the snmpwalk command I get a Timeout.

What would be the next steps to take in order to get this resolved? Any help is greatly appreciated. Thank you

[jay]

Hi

Are the servers behind a firewall? If so you may need to allow Cacti SNMP access to those hosts
Are you sure you have configured SNMP correctly on the servers?
Do the community strings match?


Cheers

Jay

[clayman2]

Hi

Are the servers behind a firewall? If so you may need to allow Cacti SNMP access to those hosts
Windows Firewall is turned off, the cacti server has iptables turned on, but I already have a rule to allow the traffic as all other servers have no issues
Are you sure you have configured SNMP correctly on the servers?
Yes I am sure SNMP is configured correctly
Do the community strings match?
Yes the community strings do match

Cheers

Jay

[jay]

Maybe an issue with agent on the servers.

I would suggest installing net-snmp on the machines and doing a local snmpwalk, if that is possible.

I assume you are adding the OID at the end of your snmwalk statement

Cheers

[bryancromwell]

Have you tried version 2 (command line -v2c)? A lot of devices don't allow version 1 snmp as well as some security devices drop version 1 snmp requests.

[clayman2]

Maybe an issue with agent on the servers.

I would suggest installing net-snmp on the machines and doing a local snmpwalk, if that is possible.
Not really an option I am looking for right now
I assume you are adding the OID at the end of your snmwalk statement
No I haven't as any other server that works fine, retrieves a bunch of information without specifying a OID
Cheers

[clayman2]

Have you tried version 2 (command line -v2c)? A lot of devices don't allow version 1 snmp as well as some security devices drop version 1 snmp requests.

With version 2 I get the following output:

Server 1 of cluster - SNMPv2-SMI::mib-2 = No Such Object available on this agent at this OID
If I specify an OID that I get from using MIB Browser for instance .1.3.6.1.2.1.2.2 I get no results

Server 2 of cluster - SNMPv2-SMI::mib-2 = No more variables left in this MIB view (It is past the end of the MIB tree)
If I specify an OID that I get from using MIB Browser for instance .1.3.6.1.2.1.2.2 I get End of MIB

Server 3 stand alone - Timeout: No response

[bryancromwell]

That tell me that you can only use version 2, You are at least getting snmp to respond. Next step to try would be to use snmpwalk to query all of the oid's and see what is supported on that device, but odds are you needed one more .# on the end of that oid for the interface number in the index.

[TheWitness]

The first two agents do not contain the oid that you are looking for the last either does not have snmp, or you have the wrong security information.

[clayman2]

Majority of my servers are Windows 2003, and some are Windows 2008. Alls I do is install SNMP on the servers, modify the agent tab and security tab under the service and everything works, except for these three servers, they are Windows 2003. So why would the others install everything and these three nothing.

The third one I have uninstalled and reinstalled SNMP, reconfigured everything and still no response.

I will try and install net-snmp on one of the servers to see if that makes a difference, but it isn't what I really want to do right now as a resolution. I will let you know of the results.

[clayman2]

I install net-snmp on the server were I get no responce by following the directions from this link http://forums.cacti.net/about26151.html

When I run the test on the server itself, I get results, so then if I run the test from the cacti server I still get a No response

[gandalf]

I install net-snmp on the server were I get no responce by following the directions from this link http://forums.cacti.net/about26151.html

When I run the test on the server itself, I get results, so then if I run the test from the cacti server I still get a No response

Either a firewall may block the traffic or the snmpd.conf may have a filter expression for valid IPs
R.

[clayman2]

Thats the weird part, there is no software installed firewall, Windows firewall is disabled, and both my client and the server are behind a hardware firewall, so that should cause any issues seeing they are both on the same network, so there is no need for the traffic to go to the hardware firewall

[gandalf]

tcpdump will help to learn, were packets are dropped (or not even send)
R.