[c-nsp] 7604 incorrent counter values on VLAN interface

[Bill Wichers]

Correct, they will wrap when you have moved more than about 4 billion bytes
(2^32) in a single 5 minute polling cycle. That works out to just over
114.5Mb/s

It's usually easiest these days to just use 64 bit counters for everything,
but with anything over essentially a 100M interface it's essential.

For the original poster: you might want to have Cacti rescan for interfaces
on the device you're watching. I've seen VLANs show bad info because Cacti
isn't polling the correct counter for the interface it thinks it's watching.
For this reason I always watch all the VLANs and the physical interface
they're bound to (on aggregator devices) so that there is a sanity check for
incorrect graphs. If all the VLANs don't add up to roughly the same as the
physical interface then you know there is a problem with one or more of the
graphs.

Also, in the web interface for Cacti, if you select the type of graph in the
lower right in the interfaces screen for a device, it will gray out all the
interfaces you're already graphing with that graph type. You can quickly
tell if you're using 32 or 64 bit counters for a given graph that way.

  -Bill


> -----Original Message-----
> From: cisco-nsp [mailto:cisco-nsp-bounces at puck.nether.net] On Behalf Of
> Mark Tinka
> Sent: Friday, May 22, 2015 2:30 PM
> To: Blake Dunlap
> Cc: cisco-nsp at puck.nether.net
> Subject: Re: [c-nsp] 7604 incorrent counter values on VLAN interface
> 
> 
> 
> On 22/May/15 20:28, Blake Dunlap wrote:
> > Are you sure this isn't a counter wrap issue? Are you using the 64 bit
> > counters or the default 32 bit in cacti?  The wrap line is right
> > around 1gig is why i ask.
> 
> I've generally found 32-bit counters to wrap around 120Mbps.
> 
> Mark.
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/