At 8:45 AM -0700 9/6/11, Randy Millsop wrote:
Hi Dennis,
Regarding your four simple suggestions, I both agree and disagree.
1) Unnecessary: I use 'Interface Name' for my Cisco devices to
achieve the result you're looking for.
Hi Randy,
Thanks for the info about 'interface name'. The reason I didn't clue
in to this is because interface names are not returned by our Cisco
switches running older images, but I see on the latest images that
they are returning 'interface names'.
3) I Disagree; Appletalk still exists on networks, and should be
monitored as each Network Admin desires. Removing a protocol
monitoring capability because someone doesn't know what it is
doesn't make it go away. People should instead educate themselves
about what is actually running on their network. This goes for all
the other protocols that many devices (such as printers) enable by
default. Perhaps, in a future version, another approach might be to
show all protocols seen on the network or in routers, and then have
both global and device options to turn off monitoring for those
protocols. Personally, I absolutely need to know everything running
on our network and keep it running smoothly - that's what I get paid
for! :-)
I think it would be a good idea to have a global setting to remove
the Appletalk feature from Set Behavior for those sites that aren't
interested in Appletalk.
4) I disagree, for similar reasons as in #3. Interface discards are
NOT meaningless. Most of my interfaces should not have discards. A
few here and there may only indicate occasional buffer shortages,
which may be acceptable if it was just caused by a max throughput
file copy, but unacceptable if it indicates that the switch/router
is not capable of handling the volume of traffic required. Excessive
discards can be a great indicator of physical line problems,
electronics going bad, or software gone crazy. There are dozens of
causes for discards, and those of us whose job it is to ensure our
network is running optimally must care.
I've opened several cases with Cisco about the high level of discards
being reported by Intermapper on all of our thousands of Cisco 3750
switches. I am told by the TAC that these counts of discards are
cosmetic and packets are not being dropped.
Every 5 minutes or so switches clear their MAC tables and briefly
flood. When this happens upstream switches receive packets that they
know are out the interface they came in on. These packets are
discarded and I suspect this causes the discard count to go up.
We ignore discards for all of our Cisco switches. It sure would be
nice to have a global setting to be able to change this behavior
globally.
Dennis
____________________________________________________________________
List archives:
http://www.mail-archive.com/intermapper-talk%40list.dartware.com/
To unsubscribe: send email to: [email protected]
