Re: portDown/portUp

[Pablo Brenner <[e-mail]>]

> From imp!spud.hyperion.com!levers.enet.dec.com!anil Tue Oct 27 08:05:18 1992
> Return-Path: <imp!spud.hyperion.com!levers.enet.dec.com!anil>
> Received: from imp.UUCP by FibHaifa.com (4.1/SMI-4.1)
> 	id AA23310; Tue, 27 Oct 92 08:05:17 IST
> Received: by imp.HellNet.org (/\==/\ Smail3.1.21.1 #21.8)
> 	id <[e-mail]>; Tue, 27 Oct 92 01:01 IST
> Received: from netcomsv.netcom.com by spud.hyperion.com (4.1/SMI-4.0)
> 	id AA04473; Mon, 26 Oct 92 12:23:13 PPE
> Received: from inet-gw-1.pa.dec.com by netcomsv.netcom.com (4.1/SMI-4.1)
> 	id AA27952; Mon, 26 Oct 92 11:10:03 PPE
> Received: by inet-gw-1.pa.dec.com; id AA15166; Mon, 26 Oct 92 11:04:15 -0800
> Received: by us1rmc.bb.dec.com; id AA05273; Mon, 26 Oct 92 13:59:58 -0500
> Message-Id: <[e-mail]>
> Received: from levers.enet; by us1rmc.enet; Mon, 26 Oct 92 14:00:09 EST
> Date: Mon, 26 Oct 92 14:00:09 EST
> From: Anil Rijsinghani <[e-mail]>
> To: [e-mail]
> Cc: [e-mail]
> Apparently-To: [e-mail], [e-mail]
> Subject: Re: portDown/portUp
> Status: R
> 
> > Agree with Paul and Anil, but if we go further, because of the
> > same reason we may need some other traps (Duplicate address found
> > for example?)
> 
>     Pablo,
> 
>     Consistent with the recommendation in RFC-1215 to define new traps
>     sparingly, and following a WG meeting prior to the advancement of
>     the FDDI MIB to PS, we had decided to eliminate traps defined
>     at the time.  (I don't believe that portDown/portUp were amongst
>     those defined in the document.)
> 
>     However from operational experience so far, we (and at least one other
>     vendor) have found that portDown traps are needed, at the least.  Two
>     customers have specifically asked for them.  I guess Duplicate
>     address found could be another reason for network problems, but
>     it doesn't seem to have been encountered in practise often enough to
>     be considered a real problem.  This is why I proposed portDown
>     at least.  The fewer the traps we define the better, I believe.
> 
>     Anil
> 

OK, I didn't want to start the discussion again, but I am in favour of traps,
even when you cannot always trust getting them (so you cannot build an SNMP
Manager that builds on traps), they really help the Network Manager.
My point is that there is no difference on implementing 1 trap or 10, and once
we define 1 trap, we should review the whole subject.



Anil, one other little issue, if we are issuing a new FDDI MIB, please
correct on the acknowledgments section my (Pablo Brenner) and Dror Shindelman
affiliation, is not Sparta but Fibronics. (I just never read this section
before :) ). Thanks.

--Pablo
--------------------------------------------------------------
Pablo Brenner
R&D Dept.
Fibronics Ltd., Matam Industrial Park, Haifa 31905, ISRAEL
phones: [phone]            Fax: [phone]
e-mail: [e-mail]   or  [e-mail]
Wireless: CX5YI/4X
---------------------------------------------------------------
=============== Usual disclaimers apply =======================