At the University of British Columbia we have several thousand
switches, mostly Cisco 3750 stacks.
In my experience with the Cisco switches we have so many interface
discards being reported that we tend to ignore interface discards
whenever and wherever they surface (by setting the "ignore interface
discards" behaviour in Intermapper). The discards happen in all
situations including under extremely light load. For whatever reason
the discards tend to occur mostly on larger switch stacks of 7, 8, or
9 switches.
I am sure there are cases where discards happen for legit reasons as
Pedro and Chris say, but in our case we spent a lot of time looking
and couldn't find the reason. It certainly wasn't anything obvious
like QoS, or rate limiting, or VLAN mismatch on an 802.1Q trunk, or
oversized frames (those get reported as real ethernet errors
(giants)), or network congestion (the discards happen under light
load). And when the discards are being reported if we do network
performance testing (iperf, ip sla) we can tell 0 packets are being
dropped in our test streams. Plus no users have ever reported
connectivity issues that correlate to the discards happening.
I've opened several cases with Cisco about this, and after Cisco
investigated was told that the reason discards were incrementing was
a "cosmetic bug". That is, the discards didn't reflect a lack of
resources in the switches or dropped packets.
So we really have no choice except to set "ignore interface
discards". I know several other large Intermapper shops that do the
same for the same reason.
PS:
As I mentioned I think the discards may correspond to flooded packets
where the recipient switch believes that the egress port is the
ingress port (based on its forwarding table) and so drops the packet,
although I have no way of confirming this. Switches are supposed to
increment their discard count in this case. But note that in this
case the discard is completely normal and doesn't indicate a network
problem, since all switches occasionally flood.
At 8:40 AM -0400 9/8/11, Romo, Chris (C) wrote:
Mismatched allowed VLANs on a trunk port will also cause InterMapper to
display "hidden" discards.
Chris Romo
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Kevin Wigle
Sent: Wednesday, September 07, 2011 10:32 PM
To: 'InterMapper Discussion'
Subject: [IM-Talk] SNMP Issues
Dear Group,
I'm sure this has been touched on before but I need a refresher on just
how
error indications on Intermapper relate to the actual hardware being
monitored.
Specifically we have a site that has two switches connected at 100 full.
Often the link between the switches will be "bubbling" away. Open
status
and it will say something like 4000 discards a minute (and higher).
When I telnet to the switch and show stats on the interface the only
errors
I see are 16 FCS errors. Not quite the 4000+ I expected to see.
So what do I trust?
How do I explain to other support personnel just what it means when IM
says
that but the actual hardware says something else?
Especially bosses - they want to see that IM is actually worth having
but if
we can't trust what we see - then what?
All opinions welcome.
regards,
Kevin
____________________________________________________________________
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]
____________________________________________________________________
List archives:
http://www.mail-archive.com/intermapper-talk%40list.dartware.com/
To unsubscribe: send email to: [email protected]
