On Wed, 2011-06-15 at 13:29 -0700, Rick Jones wrote:
> On Wed, 2011-06-15 at 14:49 -0500, Gary Gatten wrote:
> > Good info.  I don't use sflow right now but I'm sure I will in the near
> >  future.  Too bad your agent isn't hooking to 64bit counters.
> 
> Don't get too hung-up on sFlow - it is merely the messenger bringing
> wrapping counters.  RRD's response to those wrapping counters (via use
> of DERIVE) is what leads to the drops.
> 
> > So I guess the simple fix would be to sample every 10secs? Or, are u
> >  saying ANY time you encounter a wrap the derive function will fail?
> 
> I believe there is an issue any time there is a wrap while using
> DERIVE,min=0.  Notice the dips ~35 seconds apart on center in the graph
> when getting samples every 2 seconds, and then similar dips with 10
> second sampling. 

although... that does beg the question as to why those are dips and not
complete drops...

rick


_______________________________________________
Ntop mailing list
[email protected]
http://listgateway.unipi.it/mailman/listinfo/ntop

Reply via email to