I knew it was netFlow, I was curious about the mii stuff.  See detecting the
interface type, etc. is a grey art (not quite a black art, but definitely
more complex than the usual bear).  So I was curious if there was an updated
mii-diag/mii-tool that knew about these new types of interfaces...

As to the loss of data, yes, but as to the scaling, no.  If you look at the
snippet I posted, that's the only reference - it turns values > top into
DNAN, but has nothing directly to do with scaling.

I'll make the patch!

Thanks!

-----Burton

> -----Original Message-----
> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf
> Of Kouprie, Robbert
> Sent: Tuesday, January 20, 2004 3:11 AM
> To: '[EMAIL PROTECTED]'
> Subject: RE: [Ntop-dev] [PATCH] RRDPlugin - 1 Gbit is getting common...
>
>
> Hi,
>
> The 1 Gbit interface  is located on another box which runs nProbe and
> transfers NetFlow data over a regular 100 Mbit line to the ntop box. So I
> don't see how you would want to implement such a detection from
> the ntop box
> POV.
>
> Anyway, the interface is:
> 02:01.0 Ethernet controller: Sundance Technology Inc: Unknown device 1021
> (rev 0c)
>
> And mii-diag isn't showing anything useful:
> [EMAIL PROTECTED] root]# ./mii-diag eth2
> SIOCGMIIREG on eth2 failed: Invalid argument
> SIOCGMIIREG on eth2 failed: Invalid argument
> SIOCGMIIREG on eth2 failed: Invalid argument
> SIOCGMIIREG on eth2 failed: Invalid argument
> SIOCGMIIREG on eth2 failed: Invalid argument
> SIOCGMIIREG on eth2 failed: Invalid argument
> SIOCGMIIREG on eth2 failed: Invalid argument
> SIOCGMIIREG on eth2 failed: Invalid argument
> Basic registers of MII PHY #0:  ffff ffff ffff ffff ffff ffff ffff ffff.
>   No MII transceiver present!.
>   Use '--force' to view the information anyway.
>
> This is probably because i had to use an out of kernel driver for it :S
>
> Maybe we can make topValue a configurable item in the HTML
> rrdPlugin config?
> Anyway, choosing the value too low will lose data, choosing it
> too high will
> in the worst case scale some graph a little odd.
>
> -- Robbert
>
> > -----Original Message-----
> > From: Burton M. Strauss III [mailto:[EMAIL PROTECTED]
> > Sent: Monday, January 19, 2004 9:21 PM
> > To: [EMAIL PROTECTED]
> > Subject: RE: [Ntop-dev] [PATCH] RRDPlugin - 1 Gbit is getting
> > common...
> >
> >
> > Picky, picky, picky.
> >
> > it seems a reasonable change - all that's done with the value
> > is to store in
> > from the DS: line in rrd_create.c and then compare it in rrc_update.c.
> >
> >                /* make sure pdp_temp is neither too large or too small
> >                 * if any of these occur it becomes unknown ...
> >                 * sorry folks ... */
> >                if ( ! isnan(rate) &&
> >                     (( ! isnan(rrd.ds_def[i].par[DS_max_val].u_val) &&
> >                          rate >
> > rrd.ds_def[i].par[DS_max_val].u_val ) ||
> >                     ( ! isnan(rrd.ds_def[i].par[DS_min_val].u_val) &&
> >                         rate <
> > rrd.ds_def[i].par[DS_min_val].u_val ))){
> >                   pdp_new[i] = DNAN;
> >                }
> >
> > Not using 'U' is a valuable safety valve against
> > uninitialized storage,
> > however - I just wish there was a way to determine the bit rate of the
> > interface and use a reasonable value.
> >
> > What does mii-diag (ftp://ftp.scyld.com/pub/diag/mii-diag.c)
> > show for the 1
> > Gbps interface?
> >
> > -----Burton
> >
> > > -----Original Message-----
> > > From: [EMAIL PROTECTED]
> > [mailto:[EMAIL PROTECTED] Behalf
> > > Of Kouprie, Robbert
> > > Sent: Monday, January 19, 2004 8:59 AM
> > > To: '[EMAIL PROTECTED]'
> > > Subject: RE: [Ntop-dev] [PATCH] RRDPlugin - 1 Gbit is
> > getting common...
> > >
> > >
> > > I *am* using NetFlow on 1gbit ;)
> > >
> > > > -----Original Message-----
> > > > From: Burton M. Strauss III [mailto:[EMAIL PROTECTED]
> > > > Sent: Monday, January 19, 2004 3:46 PM
> > > > To: [EMAIL PROTECTED]
> > > > Subject: RE: [Ntop-dev] [PATCH] RRDPlugin - 1 Gbit is getting
> > > > common...
> > > >
> > > >
> > > > man rrd
> > > >
> > > > It suggests not using U so that rrd can perform "a basic
> > > > sanity check" on
> > > > the data.
> > > >
> > > > But I suppose 1G is reasonable for now.
> > > >
> > > > Of course, the chance that ntop can keep up with 1Gbps -
> > > > unless you're using
> > > > netFlow and heavy compression is zero...
> > > >
> > > > -----Burton
> > > >
> > > >
> > > > > -----Original Message-----
> > > > > From: [EMAIL PROTECTED]
> > > > [mailto:[EMAIL PROTECTED] Behalf
> > > > > Of Kouprie, Robbert
> > > > > Sent: Monday, January 19, 2004 7:08 AM
> > > > > To: '[EMAIL PROTECTED]'
> > > > > Subject: [Ntop-dev] [PATCH] RRDPlugin - 1 Gbit is
> > getting common...
> > > > >
> > > > >
> > > > > Use 1 Gbit as maximum line speed to compute the RRD's topValue.
> > > > > BTW: why not
> > > > > use "U" (unlimited)?
> > > > >
> > > > > -- Robbert
> > > > >
> > > > > --- rrdPlugin.c.orig    2004-01-19 11:53:28.000000000 +0100
> > > > > +++ rrdPlugin.c     2004-01-19 13:55:01.000000000 +0100
> > > > > @@ -563,7 +563,7 @@
> > > > >      int value1, value2;
> > > > >      unsigned long topValue;
> > > > >
> > > > > -    topValue = 100000000 /* 100 Mbps */;
> > > > > +    topValue = 1000000000 /* 1 Gbit/s */;
> > > > >
> > > > >      if(strncmp(key, "pkt", 3) == 0) {
> > > > >        topValue /= 8*64 /* 64 bytes is the shortest packet
> > > > we care of */;
> > > > >
> > > > >
> > > > > _______________________________________________
> > > > > Ntop-dev mailing list
> > > > > [EMAIL PROTECTED]
> > > > > http://listgateway.unipi.it/mailman/listinfo/ntop-dev
> > > > >
> > > >
> > > > _______________________________________________
> > > > Ntop-dev mailing list
> > > > [EMAIL PROTECTED]
> > > > http://listgateway.unipi.it/mailman/listinfo/ntop-dev
> > > >
> > >
> > >
> > > _______________________________________________
> > > Ntop-dev mailing list
> > > [EMAIL PROTECTED]
> > > http://listgateway.unipi.it/mailman/listinfo/ntop-dev
> > >
> >
> > _______________________________________________
> > Ntop-dev mailing list
> > [EMAIL PROTECTED]
> > http://listgateway.unipi.it/mailman/listinfo/ntop-dev
> >
>
>
> _______________________________________________
> Ntop-dev mailing list
> [EMAIL PROTECTED]
> http://listgateway.unipi.it/mailman/listinfo/ntop-dev
>

_______________________________________________
Ntop-dev mailing list
[EMAIL PROTECTED]
http://listgateway.unipi.it/mailman/listinfo/ntop-dev

Reply via email to