Sounds like you've got a buggy SNMP implementation on your switch. Try making sure you've got the latest software on your switches. You can see the snmp counters by using snmpwalk directly on most platforms ... snmpwalk -v 2c -c <YOUR COMMUNITY> <YOURHOST> ifTable
You can also walk just specific portions by replacing ifTable with the portion you're interested in, or use snmpget instead of snmpwalk to get a single specific counter. YMMV because some distributions do not setup their command line snmp stuff correctly to find the MIBs (Observium works by explicitly setting this to known working values and shipping MIB definitions)
On Fri, Feb 21, 2014 at 4:23 AM, Tom De Puysseleyr tom.depuysseleyr@belorta.be wrote:
Hi
I'm trying to setup Observium to poll all of our switches, and so far I just love it. Great job guys! I hope to convince the management to subscribe soon, but for now we'll have to do with the community version.
Let me try to explain my issue. I am polling a dual stack of Nortel Baystack 5510-24T switches using snmp v2c. All ports of unit 1 and ports 1 - 6 of unit 2 look fine. The graphics (bits/sec) for port 7 to 24 of unit 2 are strange though. Port 7 is not used, no device is connected to it at all. The average traffic according to the graphs is around 30Gbit/s. All other ports after port 7 show the exact same numbers. So first I tried to reset all port statistics on the switch. The graphs showed 0 bits/s for a few minutes, but then after let's say 30 minutes it shows a current bandwidth of 8Gbit/s. I tried deleting the device from Observium and added it again, but it seems that all counters are still there. When I look at the port using telnet, all counters are 0. I know that this switch is not in the supported devices list, but as far as I know the default snmp pollings should provide correct numbers for bandwidth statistics, no?
Any help is greatly appreciated!
Regards,
Tom De Puysseleyr
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium