Possible CPU poling issue

Ciag

I have recently started troubleshooting a issue with Orion at the moment. We have noticed that a number of devices (199 in total, mostly router boxes of varying models) are not showing any data in the CPU load and memory use graphs yet they do show all other stats. I have been looking over the devices and I have only found 1 similarity between the devices that I have checked so far, which is the Cisco IOS realease version is either (12.08) or (12.019). Am I barking up the wrong tree or is this a known issue with Orion for these releases?

 I know these releases are fairly old and they will be upgraded fairly soon but I would just like to be sure on whether this is the actual issue or not.

 If not, would you have any suggestions there are also 5 switches included the list.

 Kind Regards

Ciaran 

sedmo

Yes, I have seen this same issue and also opened a case with Solarwinds. Case #31006. 

From what I can tell, NPM first tries to poll the 1.3.6.1.4.1.9.9.109.1.1.1.1.5 oid which is not available on the older IOS versions.  It then tries to poll 1.3.6.1.4.1.9.2.1.58.0 (avgBusy5) using a snmpgetnext which returns 1.3.6.1.4.1.9.2.1.59  I can successfully poll the value from the avgBusy5 oid if I use an snmpgetnext for 1.3.6.1.4.1.9.2.1.58 or an snmpget for 1.3.6.1.4.1.9.2.1.58.0.  So either the routers are not resonding correctly to the query or they are not being queried correctly for this oid.

 

The text I submitted for my support case is listed below along with their response. 

Hi, I'm experiencing a problem with missing cpu and memory stats in Orion. This seems to be happening with cisco devices which are running versions of ios previous to 12.1. I have used a sniffer to view the polling traffic and from what I can tell Orion first tries to poll the CPUTotal5min (1.3.6.1.4.1.9.9.109.1.1.1.1.5) oid using a snmp getnext request. The router replies with 1.3.6.1.6.3.1.1.6.1.0 snmpSetSerialNo = 0. Orion then tries to poll the avgBusy5 (1.3.6.1.4.1.9.2.1.58.0) oid using another getnext request. The router replies to this with idleCount (1.3.6.1.4.1.9.2.1.59.0) whose value is generally a fairly big number. I have tested polling the avgBusy5 oid using an snmpget and the correct value is returned. My take on the problem is that Orion first tries to poll the CPUTotal5min oid but some routers with older IOS versions do not support this. Orion then tries to poll the avgBusy5 oid using a getnext request which returns the wrong oid ( I think it should be using a snmpget). Also, I do not see any poll attempts from Orion for these devices for memory utilization that I see with devices that are working correctly. This would be polls for the ciscoMemoryPoolFree (1.3.6.1.4.1.9.9.48.1.1.1.6) and ciscoMemoryPoolUsed (1.3.6.1.4.1.9.9.48.1.1.1.5) oid's. Thank you for any assistance you can provide with this issue.

 

Solarwinds response.

I verified that the way Orion ask for the OIDs cannot be changed or there is no easy way in doing this.  The best option to monitor these devices is to setup a custom poller as discussed per our phone conversation. Sorry for any inconvenience this may have caused.  I am going to close this case but if you need any further assistance with this issue please do not hesitate to give a call.