[Top] [All Lists]

Re: [nv-l] MIB data collection questions

To: nv-l@lists.us.ibm.com
Subject: Re: [nv-l] MIB data collection questions
From: justin_eagleson@bankone.com
Date: Fri, 20 May 2005 10:12:32 -0400
Delivery-date: Fri, 20 May 2005 15:13:25 +0100
Envelope-to: nv-l-archive@lists.skills-1st.co.uk
In-reply-to: <OF15CBACDC.4CDFDC15-ON85257007.000555F5-85257007.0006F30F@us.ibm.com>
Reply-to: nv-l@lists.us.ibm.com
Sender: owner-nv-l@lists.us.ibm.com
Thanks again Leslie, Channing and everyone else who reply directly and 
through the mail list. Your real-life experiences are invaluable.

Thanks and regards,

Justin Eagleson
JPMorganChase Global Technology Infrastructure
Systems and Service Management Tools
office: 614-213-9017
pager: 877-864-2056
cell: 614-226-1148

Justin Eagleson

Leslie Clark <lclark@us.ibm.com>
Sent by: owner-nv-l@lists.us.ibm.com
05/19/2005 09:23 PM
Please respond to nv-l
        To:     nv-l@lists.us.ibm.com
        Subject:        Re: [nv-l] MIB data collection questions

You did not say what platform you are running on. I can only really 
address Unix (AIX). As for scalability, I think the capacity limitation is 
more on the devices being polled than on the Netview server. How busy are 
they, and how much data can they spit out given the timeouts and retries 
you specify. A few mib variables every 5 minutes can be a lot what with 
today's devices having hundreds of interfaces. Many of the values you are 
likely to want to poll are mib expressions like bandwidth utilization, 
involving several mib variables, so that's another multiplier.   

Netview does a nice job of combining the requests into bundles, and can 
limit the size of the request, making multiple trips if necessary. This 
makes it relatively efficient. 

If you are talking hundreds of devices with hundreds of interfaces each 
and want to poll in the 5 minute range, I would not feel positive. In the 
15 minute range, I'd say maybe, if the box is big enough for everything 
else going on. For thousands of devices and only a couple of mib values 
every 15 minutes, that's probably ok too. 

I have seldom seen the snmpCollect daemon hogging the cpu unless something 
silly was configured. 

I will say, however, that I am finding it less usable these days for data 
collection on high-speed interfaces. The Counter32 datatypes it supports 
cannot handle the rollovers. In fact, at some sites the Counter64 
datatypes would not help. I see the rmon mib now has tables for reporting 
counter rollovers, so you can monitor those instead of the actual counts. 

So, let's hear it, everyone. What are you collecting, on how many 
interfaces, how often, on what hardware, and what's the bottleneck? 


Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
(248) 552-4968 Voicemail, Fax, Pager

Sent by: owner-nv-l@lists.us.ibm.com 
05/19/2005 10:06 AM 

Please respond to


[nv-l] MIB data collection questions

Does anyone care to share their experiences (good or bad) using Netview to 

poll SNMP nodes for MIB data collection? Specifically, Netview 7.1.4, and 
storing the MIB data in a flat file. How scalable was it? How many nodes 
at what interval? On what hardware was Netview running?

Any information is appreciated.


Justin Eagleson
JPMorganChase Global Technology Infrastructure
Systems and Service Management Tools
office: 614-213-9017
pager: 877-864-2056
cell: 614-226-1148

This transmission may contain information that is privileged, confidential 
and/or exempt from disclosure under applicable law. If you are not the 
intended recipient, you are hereby notified that any disclosure, copying, 
distribution, or use of the information contained herein (including any 
reliance thereon) is STRICTLY PROHIBITED. If you received this 
transmission in error, please immediately contact the sender and destroy 
the material in its entirety, whether in electronic or hard copy format. 
Thank you.

<Prev in Thread] Current Thread [Next in Thread>

Archive operated by Skills 1st Ltd

See also: The NetView Web