🚀 go-pugleaf

In article <1992Dec21.124629.19662@minerva1.bull.it> fra@minerva1.bull.it (Iarlori Francesco) writes:
>From: fra@minerva1.bull.it (Iarlori Francesco)
>Subject: ShareWare NMS machine for SNMP.
>Date: Mon, 21 Dec 92 12:46:29 GMT
>
>	Someone knows where I can find an NMS shareware version to test
>	my Agent Machine in an SNMP Network ?

Be warned that bugs lurk in many of the unsupported, freely available 
SNMP engines.

Make sure that your code properly ASN.1/BER encodes counters/gauges/timeticks
(i.e. the unsigned things) properly when the 32nd bit (the sign bit on 32-
bit integers) is set.  It should take *5* octets length do this, not 4 as in 
some code.

Also check that get next really works.  It should be able to take *any* oid, 
whether or not that oid represents a variable, and finds the sucessor (or 
end of the mib.)

Also check that your strings (e.g. sysDescr) are real NVT, i.e. that if 
there are multiple lines that the lines are separated by CR LF, not merely 
LF.  And don't send any trailing nulls.

			--karl--

In article <karl.37.725592903@empirical.com>, karl@empirical.com (Karl Auerbach)
says:
>
>Also check that get next really works.  It should be able to take *any* oid,
>whether or not that oid represents a variable, and finds the sucessor (or
>end of the mib.)
>
I wrote some code to get the ARP tables using SNMP. I start by doing a get-next
on mgmt_mib_ip_ipNetToMediaTable_ipNetToMediaEntry_ipNetToMediaPhysAddress, and
I look at the type of the returned variable. If it is STRING I know the unit
supports MIB-II so I use those variables. Otherwise I assume it does not and I
use the MIB-I at table.

I tested this against various boxes (nodes, routers, etc), and found that the
Cabletron MRXI 10Base-T hub, which supports MIB-I, appears to fail in its
treatment of get-next. It returns NoSuchVariable for the get-next I outlined
above.

Anybody else seen this? Does Cabletron know about it?

Tim.