graph dips every 16 to 17 minutes (cisco 6509)

[Tenaku]

I am pulling snmp data from a cisco 6509 to graph a standard 'bits per second' on a few of the interfaces. Every 16 to 17 minutes the graph shows the data rate dropping to almost nothing then resuming as normal one minute later.

All ports on this particular switch show this, but I am graphing 50 other switches which all behave normally. Is it an issue with cisco 6509s? I only have one 6500 series, so I can't test against a different switch.

Any help would be appreciated.

running cacti v0.8.7b (via cactiEZ)

[Howie]

What you could do though, is check with a different measurement software... then you would know if it's SNMP on the switch, or something in Cacti.

Try PRTG (it's free for 10 interfaces) on the same interface, and see if *it* drops like that too. You can easily adjust the polling time on PRTG to be quicker, and see if that makes a difference too.

At least then you know if it's a Cacti problem or not.

[TheWitness]

I suspect a cyclic traffic pattern. You could also use Realtime. That tells a good story.

TheWitness

[billibarou]

I was having the exact same problem on both my 6500 and 7206 router. I switched to SNMP version 2 and cacti 64 bit graphs. I backed up then deleted my existing data and rebuilt the graphs and now they are smooth and not choppy.