SNMP results not valid

[andrew123456]

I've been running cacti for about 2 months without a problem on the same server as groundworks (using a few of groundwork's libs). However a few days ago I was asked to have groundworks monitor an additional network. So I plugged in another cable, set up a few static routes and did a couple firewall configs.

So now my SNMP in cacti is turning up blank, with the log saying things like this:

Code: Select all

07/06/2010 02:15:04 PM - SYSTEM STATS: Time:3.2073 Method:cmd.php Processes:1 Threads:N/A Hosts:15 HostsPerProcess:15 DataSources:217 RRDsProcessed:109
07/06/2010 02:15:03 PM - CMDPHP: Poller[0] Host[16] DS[726] WARNING: Result from SNMP not valid. Partial Result: U
07/06/2010 02:15:03 PM - CMDPHP: Poller[0] WARNING: SNMP Get Timeout for Host:'172.17.1.230', and OID:'.1.3.6.1.2.1.2.2.1.10.3000'
07/06/2010 02:15:03 PM - CMDPHP: Poller[0] Host[16] DS[726] WARNING: Result from SNMP not valid. Partial Result: U
07/06/2010 02:15:03 PM - CMDPHP: Poller[0] WARNING: SNMP Get Timeout for Host:'172.17.1.230', and OID:'.1.3.6.1.2.1.2.2.1.16.3000'
07/06/2010 02:15:02 PM - CMDPHP: Poller[0] Host[10] DS[346] WARNING: Result from SNMP not valid. Partial Result: U
07/06/2010 02:15:02 PM - CMDPHP: Poller[0] WARNING: SNMP Get Timeout for Host:'172.17.1.216', and OID:'.1.3.6.1.2.1.2.2.1.16.3000'
07/06/2010 02:15:02 PM - CMDPHP: Poller[0] Host[10] DS[346] WARNING: Result from SNMP not valid. Partial Result: U
07/06/2010 02:15:02 PM - CMDPHP: Poller[0] WARNING: SNMP Get Timeout for Host:'172.17.1.216', and OID:'.1.3.6.1.2.1.2.2.1.10.3000' 

Which, to me, looks like a firewall issue, but when I do a verbose query of a device I get results, or if I do a manual snmpwalk from that server it is able to return values.

Any ideas why it wouldn't get values normally?

[gandalf]

When walking manually, which timeout do you apply? You may want to increase the SNMP timeout on the host edit screen
R.

[andrew123456]

Thanks, but I just go with the default timeout. The response is quick too, as the switch I'm testing with is on the same LAN (in the same room, actually, with 1 hop to get there) So I don't understand how it could timeout.

[andrew123456]

okay, a slight update; when the log said this:

Code: Select all

07/07/2010 09:35:02 AM - CMDPHP: Poller[0] WARNING: SNMP GetNext Timeout for Host:'172.17.1.212', and OID:'.1.3' 

i ran this:

Code: Select all

snmpget 172.17.1.212 -v 1 -c DellSNMP .1.3

and got this:

Code: Select all

Error in packet
Reason: (noSuchName) There is no such variable name in this MIB.
Failed object: SNMPv2-SMI::org

Could someone enlighten me?

[BSOD2600]

You should try snmpwalk .1.3.6.1.2.1.2.2.1.10 instead. If that doesn't work, then the device doesn't support that OID.

[andrew123456]

that one didn't work either, saying

Code: Select all

Error in packet
Reason: (noSuchName) There is no such variable name in this MIB.
Failed object: IF-MIB::ifInOctets

but the real question here is how could it work fine before, and not now?

i have another switch, on the same LAN, and the same type of switch (Dell Powerconnect 5448) that is getting data just fine now, but I haven't done anything special.

I realize I am giving very little information here, so if some other config dump or something would be helpful let me know.

[BSOD2600]

well does ANY snmp work against this problem switch?

[andrew123456]

That's just the problem - I can perform most SNMP queries against the switch manually, but cacti itself tells me there is a timeout.

I should mention that every switch gets a timeout except for one. And I have no idea why that one works and the others don't
(they're the same type of switch on the same network with the same configs)

[BSOD2600]

what cacti/spine versions? you should not be using 0.8.7f.
what poller you using?

[andrew123456]

I am using .0.8.7f, and I haven't updated the version for some time if I recall. I'll drop down a version and see if it helps.

and I'm using the cmd.php poller, as it was working well for me at the time, so I saw no reason to switch to spine. I can give that a shot too, I suppose.

[BSOD2600]

0.8.7f is flawed. either go back to 0.8.7e or 0.8.7g beta.

[andrew123456]

okay so i've blown the entire cacti out of the water (removed all the files, dropped the mysql table) and loaded up 0.8.7e.

I STILL have this problem, so I don't know. Does this mean it's not actually a cacti issue then?

[TheWitness]

The problem here is that cacti is using the OID .1.3 as a getnext to perform the snmp ping. The device is responding, but with the message that no such object exists.

This is in-fact a good ping response. So, the snmp ping in lib/ping.php should be enhanced now that Cacti can, in fact, detect a "real" timeout vs. an error response.

In the mean time, if you can find an OID that "all" your devices respond wll to using a "getnext" operation, edit lib/ping.php and place it into the ping_snmp() function and you will be back on the air.

Also, please log a bug at http://bugs.cacti.net and add your comments there and patch there. Please use 0.8.7g.

TheWitness

[andrew123456]

Thanks for the tip, I finally got around to trying it, but that doesn't seem to do it. I changed the OID it was using for pings and it still isn't co-operating.

The problem is with cacti dealing with the results. I will get something like this in my log:

Code: Select all

07/21/2010 09:00:02 AM - CMDPHP: Poller[0] Host[2] DS[58] WARNING: Result from SNMP not valid. Partial Result: U
07/21/2010 09:00:02 AM - CMDPHP: Poller[0] WARNING: SNMP Get Timeout for Host:'172.17.1.213', and OID:'..1.3.6.1.2.1.2.2.1.10.42' 

and running this from a terminal on that machine gives me this:

Code: Select all

snmpget -v 1 -c DellSNMP 172.17.1.213 .1.3.6.1.2.1.2.2.1.10.42
IF-MIB::ifOutOctets.51 = Counter32: 2778080737

this is pretty frustrating. anyone have any more suggestions for me?

[blueman176]

I have exactly the same problem. "SNMP get timout" whereas the command line works well.