Re: Livingston SNMP bugs

root (root@sasami.anime.net)
Mon, 2 Oct 1995 12:21:57 -0700 (PDT)

> Please don't think I'm trying to argue with you. All that I am saying
> is that it is perfectly legal for an SNMP implementation to return 0
> for a value.
> On systems I have worked with where the interface is externally
> clocked, I have done both; returned 0 and returned the value that the
> customer entered. There doesn't seem to be a right way to do this,
> and I have received complaints no matter how I did it.

The Livingston docs state that the speed is simply a place holder, because
the Livingstons use external clocking. So what's to stop you from setting
the port speed to 0 if you want? It shouldn't hurt anything. However, it
*SHOULD* be returned in ifSpeed.

So _either way_, there should be _no arguments_. If the customer wants 0,
they enter 0. If they want 1536000 they enter 1536000. Simple. Problem
is, the Livingston doesn't pass it in SNMP which is *incorrect* IMHO.

And returning 0 for the speed of ether0 is totally bogus, IMHO ;)

> I guess my question then is this: If you know what the interface speed
> is and you have to enter it into the PM, why can't you also plug it
> into your software? This is certainly going to fix things a lot
> quicker than Livingston is going to respond to your problem. Again,
> I'm not trying to be argumentative here, just realistic.

Because the software doesnt' allow it (currently). It depends on the
device to return the correct value for ifSpeed. Ciscos do it.
Morningstars do it. CMU-SNMP does it. Just about everything *except*
Livingston does it.

Why allow plugging in of the interface speed at all, if the Livingston
isn't going to do anything with it at all anyway? Seems rather stupid to me.

-Dan