Cisco Routers return "SNMP Error" no matter what I do.

[asynakie]

I'm trying to add all of our Cisco routers to Cacti. In the past we used MRTG to monitor bandwidth and it would pull data fine. When I try to add a new device in Cacti for a particular router all it gives me is "SNMP Error" in RED. I know the community name is correct. I even went back into the config files from MRTG and verified the community names they were generated with. I'm using version 1, but I also tried version 2 to no avail. I can add switches all day with no problem and it pulls in all my information. When I try the routers however, I can't get anywhere. Is there something simple I'm missing?

I'm currently pulling data from all these routers in MRTG just fine using their community names. Why can't I do the same in Cacti?

[gandalf]

Due to performance reasons, Cacti runs an availability test. It snmpget's .1.3.6.1.2.1.1.1.0 for the webpage (and some more) to test device availability. In case the device does not respond (in time), it is considered as "down"
R.

[artagel]

Do you have an snmp access list? if you aren't running on the same server as MRTG then it's plausible a secure configuration that uses an snmp access-list would only have the MRTG server on it, not the new one.
-Dan

[asynakie]

Do you have an snmp access list? if you aren't running on the same server as MRTG then it's plausible a secure configuration that uses an snmp access-list would only have the MRTG server on it, not the new one.
-Dan

That might be possible. I will ask my boss on Monday to see if that's the case. Every switch (Nortel Baystack) in our organization worked just fine in Cacti, but it would make sense that the routers might have an access list set up. I'll find out and post back. Thank you.

[asynakie]

Well I changed over the new server to the same IP address of the old server and I still couldn't get any information from the routers. I create a new device and I'm greeted with SNMP Error in red at the top. If i click on debugging for "Verbose Query" for SNMP - Interface Statistics I get the following. Any other ideas?

Code: Select all

+ Running data query [1].
+ Found type = '3' [snmp query].
+ Found data query XML file at '/var/www/html/cacti/resource/snmp_queries/interface.xml'
+ XML file parsed ok.
+ Executing SNMP walk for list of indexes @ '.1.3.6.1.2.1.2.2.1.1'
+ No SNMP data returned
+ Found data query XML file at '/var/www/html/cacti/resource/snmp_queries/interface.xml'
+ Found data query XML file at '/var/www/html/cacti/resource/snmp_queries/interface.xml'
+ Found data query XML file at '/var/www/html/cacti/resource/snmp_queries/interface.xml'

Interestingly I logged into the old server after I changed the IP addresses back and I tried to manually snmpwalk the router.

snmpwalk -v 1 -c public 10.4.1.1

and I received Timeout: No Response from 10.4.1.1

However, my graphs are still updating in MRTG. Arg, this makes no sense.

edit: I've also modified the snmpd.conf file by changing this line view systemonly included .1.3.6.1.2.1. <<< new value as I've found in the forums here after searching. That did not fix the problem.

[artagel]

If you receive an snmpwalk timeout then it can't find the server. You'll have to do some troubleshooting to see why this is. Make sure you can ping the devices. Perhaps do a wireshark to see if mrtg is really running properly and connecting to the same machines via snmp. Hard to troubleshoot this remotely through a forum without access, but I believe once you get the timeout resolved you'll be good to go.
-Dan

[asynakie]

If you receive an snmpwalk timeout then it can't find the server. You'll have to do some troubleshooting to see why this is. Make sure you can ping the devices. Perhaps do a wireshark to see if mrtg is really running properly and connecting to the same machines via snmp. Hard to troubleshoot this remotely through a forum without access, but I believe once you get the timeout resolved you'll be good to go.
-Dan

I can ping all the devices no problem. I installed wireshark and there is a ton of SNMP traffic going between my server and all the other devices. When I filter out just traffic to and from the router I'm looking at, I see the SNMP get-requests come through from the server to the router. There's never any traffic coming back the other direction. Below I'm attaching a screenshot of the packet analysis from wireshark.



AT&T is in control of our routers and there are access lists in place. I turned off the port the old server is connected to and changed the IP address of my new server with Cacti to be the same as the old. I still couldn't get a response from the router, and all the SNMP traffic is the same.

[gandalf]

I can ping all the devices no problem. I installed wireshark and there is a ton of SNMP traffic going between my server and all the other devices. When I filter out just traffic to and from the router I'm looking at, I see the SNMP get-requests come through from the server to the router. There's never any traffic coming back the other direction.

Then, this is very likely not a Cacti issue as it send SNMP requests. Only thing I can imagine is the use of snmpbulkwalk. Some devices don't like that. A valid test is to switch to SNMP V1 as this does not know nothing about bulkwalks. If SNMP works then, it is a bulkwalk issue. If this does not change anything, it is very likely a target or firewall issue
R.

[asynakie]

I can ping all the devices no problem. I installed wireshark and there is a ton of SNMP traffic going between my server and all the other devices. When I filter out just traffic to and from the router I'm looking at, I see the SNMP get-requests come through from the server to the router. There's never any traffic coming back the other direction.

Then, this is very likely not a Cacti issue as it send SNMP requests. Only thing I can imagine is the use of snmpbulkwalk. Some devices don't like that. A valid test is to switch to SNMP V1 as this does not know nothing about bulkwalks. If SNMP works then, it is a bulkwalk issue. If this does not change anything, it is very likely a target or firewall issue
R.

Thanks for the reply. I've tried it both as v1 and v2 and have the same results. I know the server is working and I'm collecting data from everything else I'm monitoring. It's just for some reason I can't get to the router. It's strange that even when I change the new server's IP address to be the same as the existing server, that I still can't get any data from the routers. The old server is pulling data using MRTG without any problems using that address. I would think that changing the IP addresses around would rule out a firewall problem as well, no?

[gandalf]

In case the working MRTG graphs are located on a different server (with a different IP than the current Cacti server):
it may happen, that the target device has a list of allowed SNMP query hosts
R.

[asynakie]

In case the working MRTG graphs are located on a different server (with a different IP than the current Cacti server):
it may happen, that the target device has a list of allowed SNMP query hosts
R.

I believe there is, however I have switched off the network port on the MRTG server and assigned the current Cacti server the old server's IP address. That way even if there was a list, the new server should be allowed to query SNMP from that router. This still did not work though, so I believe something else must be at play.