nv-l
[Top] [All Lists]

Re: [nv-l] Bandwidth Util

To: nv-l@lists.us.ibm.com
Subject: Re: [nv-l] Bandwidth Util
From: Leslie Clark <lclark@us.ibm.com>
Date: Sun, 27 Feb 2005 12:59:00 -0500
Delivery-date: Sun, 27 Feb 2005 17:59:27 +0000
Envelope-to: nv-l-archive@lists.skills-1st.co.uk
In-reply-to: <20050224080809.80425.qmail@web60407.mail.yahoo.com>
Reply-to: nv-l@lists.us.ibm.com
Sender: owner-nv-l@lists.us.ibm.com

You will have to do a comparison with the device to determine whether the numbers are accurate or not.

To understand what mib variables are being pulled for those expressions, take a look at /usr/OV/conf/mibExpr.conf. It is pulling ifInOctets and/or ifOutOctets, times 8 to get bits, and dividing by ifSpeed, which is in bits/sec. Netview is still using 32-bit variables to pull ifInOctets and ifOutOctets. Your devices may have Gig-speed interfaces that store counters in 64-bit variables. Those variables are counters that keep increasing until they reach the max of the hardware, then roll back to zero.   Netview will automatically drop a value when it finds that a value is lower than the last
value it found, understanding that it has wrapped.  

This used to work just fine. But with the new higher-speed interfaces, it is possible to wrap the counter many times between polls. You have no way of knowing how many times it has wrapped. And if Netview only sees part of the variable, there is no way to guarantee that it is anywhere near the truth. The number may just look lower because the low-order digits are lower. Or it may appear to have grown, but it has actually wrapped with the low-order digits ending up higher.

You should do a sanity-check on one of your devices. Use the mib browser to watch the values of ifInOctets or ifOutOctets for one of your intefaces, and do the math along with Netview. At the same time,  telnet into the device and  show the values for traffic on that same interface. See if they match up. This will tell you whether or not snmpCollect is capable of monitoring this device. If you can go several polling intervals without wrapping, either for real, or as reported by Netview, then you can monitor that interface with snmpCollect.

Cordially,

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



Bhayeti <bhayeti@yahoo.com>
Sent by: owner-nv-l@lists.us.ibm.com

02/24/2005 03:08 AM
Please respond to
nv-l

To
nv-l@lists.us.ibm.com
cc
Subject
[nv-l] Bandwidth Util





Folks,

I'm using the built in "BandwidthUtilHdx' collection
in SNV 7.1.4 on AIX. Could someone pls assist me to
understand what the raw value collected actually
means. Do I need to perform an additional calculation
to make sence of the number?

Regards,

B


                                 
__________________________________
Do you Yahoo!?
Yahoo! Mail - Easier than ever with enhanced search. Learn more.
http://info.mail.yahoo.com/mail_250

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

Archive operated by Skills 1st Ltd

See also: The NetView Web