Path: utzoo!utgpu!watserv1!watmath!att!news.cs.indiana.edu!samsung!zaphod.mps.ohio-state.edu!usc!ucsd!ucbvax!ANDREW.CMU.EDU!tjh+
From: tjh+@ANDREW.CMU.EDU (Tom Holodnik)
Newsgroups: comp.protocols.appletalk
Subject: Re: SNMP Experience in AppleTalk/IP
Message-ID: <EbPwyu_00WCoM5xZYU@andrew.cmu.edu>
Date: 19 Dec 90 20:38:18 GMT
References: <1221@wcc.oz>
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 26


> Excerpts from AppleTalk: 19-Dec-90 SNMP Experience in AppleTal.. Tom
> Evans@apple.com (460)

> What SNMP tools are being used out there to help administer AppleTalk
> internets, particularly those connected to IP networks.


	Last summer, I wrote a quick and hairy SNMP monitor for our population
of FastPaths, based on Steve Waldbusser's SNMP library, running on a
UNIX host. While it monitors about 50 variables, it doesn't check tables
for consistency (something I'd like to do someday soon). 
	It's still in a pretty rough state (I'm not sure how well it would pass
through lint), but it works enough to indicate several things. The most
frequently incremented counter is the counter for ALAP collisions -
those instances where the gateway sends an RTS packet, but gets no CTS
in return (presumably, it collided with someone else's transmission). 
	The other most significant counter indicates severe faults- the
alapNoData counter indicates that the gateway received an RTS, sent a
CTS, but got no data from the user's machine. Typically, this means that
some patch panel somewhere is flakey, or that there's a problem with one
of the wiring concentrators. 
	If you'd like more details, let me know. It's still very much a work in
progress, but it has shown some things I thought were interesting.

-tom