On the surface this would seem to be very similar if not identical to the
request I posted in the past week for an "ignore link state transitions"
flag.

My situation is switches with dynamic ports that may be up or down at any
given point in time, but can dynamically change at any time without
indicating a problem.

On closer thought, this particular request is a bit different in that Joe
probably does not want to ignore link state transitions (as in a failed or
flapping port) but only an original/initial port down state -- but that
assumes that he might be interested in the state of the storage device. 

This would seem to beg an underlying question(s) of what network events or
changes in operation are significant and to whom within a organization are
they significant (or to whom the change should be communicated).  In this
context of this discussion these questions are rhetorical as every
environment is different.  The trick is to build into IM the operational
flexibility such that it can adjust or be configured to operate usefully
across a plethora of situations many of which are unique. 

(I would humbly suggest that good design of an active/standby interface pair
keeps both "UP" with one being inactive -- ideally with heat beat.  How else
would one know that there that the standby is "hot" and has not suffered a
link fault (bad transmitter, receiver, cable strand) while not in use.  But
that is a hardware vendor issue not one that Dartware can solve, but it is
limitations such as these to which IM must adapt.)

In response to Janice's response to Joe's original note.  Janice's
suggestions works well for a small implementation without too many
transitions, and without too many different corner case situations to
manage, but I would suggest that it doesn't scale well in complex
environments.  As a management strategy I am trying to maximize actionable
information IM can deliver while at the same time trying to minimize
non-actionable items (cries of wolf / false positives) that are produced and
at the same time trying to minimize the cost of build out and "futz".

Anyway, some random thoughts on a Friday afternoon.

-Dave Cooley
Colby College

-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Joe Drees
Sent: Thursday, December 03, 2009 3:05 PM
To: [email protected]
Subject: [IM-Talk] Ignore Port down???

Have a new situation in our monitoring.  The Storage Group has put up an
iSCSI environment attached to switches that we monitor (Nortel 5510¹s and I
use the SNMP Traffic Probe).

But the environment uses ³standby interfaces².  They are down until a
failure of a primary requires their use.

So, if I want to monitor traffic, and potential traffic, for the
environment, I need to be displaying ³Down² interfaces on the switch,
thereby having the switch showing Alarm Status.

Anyone know of a way around this, or have any different ideas to tackle
this?


____________________________________________________________________
List archives: 
http://www.mail-archive.com/intermapper-talk%40list.dartware.com/
To unsubscribe: send email to: [email protected]

____________________________________________________________________
List archives:
http://www.mail-archive.com/intermapper-talk%40list.dartware.com/
To unsubscribe: send email to: [email protected]

Reply via email to