How many routes are in the route table? I wouldn't expect high CPU on the
Cat6500 unless you had at least a couple of thousand routes in the table.
Could also me the ARP cache. I had the same problem on Cisco 7507 that had a
34,000 entry ARP cache. When NetView did config poll, CPU would go to 100%
on that guy for 40+ minutes, due to the fact that SNMP V1 is being used and
that each entry is pulled, one at a time.
On the Cisco gear, you can configure an snmp view to force it not to forward
the ARP table. Should be able to do the same thing with the route table.
Here are the Cisco commands to disable it from sending it's ARP cache:
snmp-server view noarp iso included
snmp-server view noarp mib-2.3 excluded
snmp-server community public view noarp RO
disabling route table would be similar. You'd just need to dig thru MIB2 and
figure out which variable(s) are used to pull the route table.
__________________________
Thanks,
Sean Davidson
Sr. Network Systems Engineer
Publix Super Markets, Inc.
P.O. Box 32015
Lakeland, Fl. 33802-2015
Email - sean.davidson@publix.com
Voice - (863) 686-8754 x6889
Fax - (863)616-5895
-----Original Message-----
From: Leslie Clark [mailto:lclark@US.IBM.COM]
Sent: Tuesday, January 18, 2000 2:29 AM
To: NV-L@UCSBVM.ucsb.edu
Subject: Re: How does Netview do a Discovery Poll
There has recently been quite a bit of discussion on this topic. Netview
does pull the routing table during new node discovery and demand polls.
Also I believe during new node discovery. And this is known to have an
impact on Cisco devices. 'At the current release' you have a couple of
options to improve the situation. You can limit the number of routing table
entries pulled. This is done in Options...SNMP Configuration and can be
done on a per-device basis, or by network, or as the default. If you don't
set it it defaults to 800. You can set that low (0?) once your map is
drawn.
Also once your network is discovered, you have the
option of turning off configuration polling, which happens by default once
per day, and more or less any time you restart netmon. You also have the
option of turning off new node discovery, or you can set the cycle of
new node discovery to a longer period. By default, new node discovery
is done frequently at first, then tapers off as fewer new things are found.
If you do a demand poll on a node, the output includes information on
the schedule of the next new node discovery poll and configuration poll,
so you can get an idea there. I understand that you can also configure the
device itself to not respond to requests for information that are too
taxing.
Cordially,
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
(248) 552-4968 Voicemail, Fax, Pager
---------------------- Forwarded by Leslie Clark/Southfield/IBM on
01/18/2000 02:21 AM ---------------------------
Graeme Nelson <GNELSON@MACQUARIE.COM.AU>@UCSBVM.UCSB.EDU> on 01/18/2000
12:03:53 AM
Please respond to Discussion of IBM NetView and POLYCENTER Manager on
NetView <NV-L@UCSBVM.UCSB.EDU>
Sent by: Discussion of IBM NetView and POLYCENTER Manager on NetView
<NV-L@UCSBVM.UCSB.EDU>
To: NV-L@UCSBVM.UCSB.EDU
cc:
Subject: How does Netview do a Discovery Poll
Our Network group has raised a concern that the way NV does its Discovery
Poll. They believe NV is causing a high CPU until on MSM card in Cat6500.
What they are saying is happening is NV is requesting individual elements
of the route table rather than the table in one go. What I need to know is
which SNMP MIB variables does it retrieve during a Demand Poll, a Discovery
Poll and a Configuration Poll?
Our setup is:
Netview 5.1.1
Solaris 2.6
The 6500 switches are running version 5.3(4)CSX software and the MSMs are
running 12.0(4a)WX5(11).
Thanks
Graeme Nelson
Technical Specialist (Tivoli)
Macquarie Bank Ltd
|