nv-l

RE: [nv-l] (NV 7.1.2) Bandwidth Utilization Collects

2002-09-27 12:01:54
Subject: RE: [nv-l] (NV 7.1.2) Bandwidth Utilization Collects
From: "Shipley, Rob" <rshipley AT state.mt DOT us>
To: "'David_Moon AT lincolnelectric DOT com'" <David_Moon AT lincolnelectric DOT com>, "'nv-l AT lists.tivoli DOT com'" <nv-l AT lists.tivoli DOT com>
Date: Fri, 27 Sep 2002 10:01:54 -0600
Here are some suggestions:
1. ifInOctets & ifOutOctets are 32bit Counters. They increment to 4294967296
and then reset to 0. (This is what the snmpgets are showing) Netview
understands this and calculates delta. All you should have to do is multiply
the (snmpCollect value x8 /ifSpeed) x100. This should give you utilization.
2. Are you polling an interface on a switch that is more than 100 meg and
beyond 5 minutes? If so, ifInOctets and ifOutOctets will not report the
utilization correctly. The reason for this is because they are 32bit
counters. ( 32bit counters roll over too quickly.) You can possibly poll
every minute and then you can get utilization up to 575meg. You can also use
a 64bit counter. It depends on the manufacturer as to what OID to use. 
3. For switches, it is most likely full-duplex, so do NOT add ifInOctets and
ifOutOctets (inOctets + outOctets *8) /ifSpeed *100. This is for
half-duplexing. For full duplex: (inOctets "OR" outOctets *8 /ifSpeed) *100.

4. Check out
http://www.cisco.com/warp/public/477/SNMP/calculate_bandwidth_snmp.html
for more details.

-----Original Message-----
From: Dejan Mijailovic [mailto:dejan.mijailovic AT inquent DOT com]
Sent: Friday, September 27, 2002 7:21 AM
To: David_Moon AT lincolnelectric DOT com; nv-l AT lists.tivoli DOT com
Subject: RE: [nv-l] (NV 7.1.2) Bandwidth Utilization Collects


David.

They don't need to reset.

What you see is a TOTAL amount of bits In or Out of Interface.
What you need is real counter. If you do "snmpget" you will get the BIG
number, but snmpCollect should
collect bits (or octets) per second. It always does it if the MIB object you
collecting is  a (COUNTER)
and in you case, it is.
Check your mibExpr.conf file. Don't forget, it is a postFix math.

I think you are doing something wrong there.

Regards

Dejan



-----Original Message-----
From: David_Moon AT lincolnelectric DOT com
[mailto:David_Moon AT lincolnelectric DOT com]
Sent: Thursday, September 26, 2002 4:57 PM
To: Dejan Mijailovic; nv-l AT lists.tivoli DOT com
Subject: RE: [nv-l] (NV 7.1.2) Bandwidth Utilization Collects



Dejan and list,

Thanks for the information. The information you provided on the expression
got me to dig into this problem a little more and I think I have found the
dreaded monkey wrench! I mentioned earlier that my bandwidth utilization
numbers are very high. So I decided to manually pull the ifInOctets and
ifSpeed on one of my switches. I thought that I could then manually
calculate the expression to see what I would get. Well, lo and behold I
discovered that our ifInOctets and ifOutOctets counters do not reset. Here
is an example of what I am talking about...

systst:/usr/OV/databases/snmpCollect>snmpget 8540A.lecis.com
.1.3.6.1.2.1.2.2.1.10.5
interfaces.ifTable.ifEntry.ifInOctets.5 : Counter: 3533604219
systst:/usr/OV/databases/snmpCollect>snmpget 8540A.lecis.com
.1.3.6.1.2.1.2.2.1.10.5
interfaces.ifTable.ifEntry.ifInOctets.5 : Counter: 3533682103
systst:/usr/OV/databases/snmpCollect>snmpget 8540A.lecis.com
.1.3.6.1.2.1.2.2.1.10.5
interfaces.ifTable.ifEntry.ifInOctets.5 : Counter: 3533690318
systst:/usr/OV/databases/snmpCollect>snmpget 8540A.lecis.com
.1.3.6.1.2.1.2.2.1.10.5
interfaces.ifTable.ifEntry.ifInOctets.5 : Counter: 3533700776
systst:/usr/OV/databases/snmpCollect>snmpget 8540A.lecis.com
.1.3.6.1.2.1.2.2.1.10.5
interfaces.ifTable.ifEntry.ifInOctets.5 : Counter: 3533712130
systst:/usr/OV/databases/snmpCollect>snmpget 8540A.lecis.com
.1.3.6.1.2.1.2.2.1.10.5
interfaces.ifTable.ifEntry.ifInOctets.5 : Counter: 3533713363
systst:/usr/OV/databases/snmpCollect>snmpget 8540A.lecis.com
.1.3.6.1.2.1.2.2.1.10.5
interfaces.ifTable.ifEntry.ifInOctets.5 : Counter: 3533741144
systst:/usr/OV/databases/snmpCollect>snmpget 8540A.lecis.com
.1.3.6.1.2.1.2.2.1.10.5
interfaces.ifTable.ifEntry.ifInOctets.5 : Counter: 3533769809
systst:/usr/OV/databases/snmpCollect>snmpget 8540A.lecis.com
.1.3.6.1.2.1.2.2.1.10.5
interfaces.ifTable.ifEntry.ifInOctets.5 : Counter: 3533786943
systst:/usr/OV/databases/snmpCollect>snmpget 8540A.lecis.com
.1.3.6.1.2.1.2.2.1.10.5
interfaces.ifTable.ifEntry.ifInOctets.5 : Counter: 3533799415
systst:/usr/OV/databases/snmpCollect>snmpget 8540A.lecis.com
.1.3.6.1.2.1.2.2.1.10.5
interfaces.ifTable.ifEntry.ifInOctets.5 : Counter: 3533845173
systst:/usr/OV/databases/snmpCollect>snmpget 8540A.lecis.com
.1.3.6.1.2.1.2.2.1.10.5
interfaces.ifTable.ifEntry.ifInOctets.5 : Counter: 3536594117
systst:/usr/OV/databases/snmpCollect>snmpget 8540A.lecis.com
.1.3.6.1.2.1.2.2.1.10.5
interfaces.ifTable.ifEntry.ifInOctets.5 : Counter: 3536826662
systst:/usr/OV/databases/snmpCollect>snmpget 8540A.lecis.com
.1.3.6.1.2.1.2.2.1.10.5
interfaces.ifTable.ifEntry.ifInOctets.5 : Counter: 3537038764
systst:/usr/OV/databases/snmpCollect>snmpget 8540A.lecis.com
.1.3.6.1.2.1.2.2.1.10.5
interfaces.ifTable.ifEntry.ifInOctets.5 : Counter: 3537333694
systst:/usr/OV/databases/snmpCollect>snmpget 8540A.lecis.com
.1.3.6.1.2.1.2.2.1.10.5
interfaces.ifTable.ifEntry.ifInOctets.5 : Counter: 3537662664
systst:/usr/OV/databases/snmpCollect>snmpget 8540A.lecis.com
.1.3.6.1.2.1.2.2.1.10.5
interfaces.ifTable.ifEntry.ifInOctets.5 : Counter: 3537873300
systst:/usr/OV/databases/snmpCollect>snmpget 8540A.lecis.com
.1.3.6.1.2.1.2.2.1.10.5
interfaces.ifTable.ifEntry.ifInOctets.5 : Counter: 3538178381
systst:/usr/OV/databases/snmpCollect>snmpget 8540A.lecis.com
.1.3.6.1.2.1.2.2.1.10.5
interfaces.ifTable.ifEntry.ifInOctets.5 : Counter: 3538485680
systst:/usr/OV/databases/snmpCollect>snmpget 8540A.lecis.com
.1.3.6.1.2.1.2.2.1.10.5
interfaces.ifTable.ifEntry.ifInOctets.5 : Counter: 3538767740

Now in my pea-sized mind, I would think that these counters should reset or
at least that Netview would somehow calcuate the 'delta' value between two
points in time. However, I do not believe this is happening based on what I
have seen so far. Has anyone run into this before? Do I need to have the
network folks configure their devices in a certain way.

Thanks for the input,
Dave Moon
Systems Architect
Lincoln Electric Company




                    "Dejan Mijailovic"
                    <[email protected]       To:
<David_Moon AT lincolnelectric DOT com>,
                    quent.com>                  <nv-l AT lists.tivoli DOT com>
                                               cc:
                    09/26/02 03:29 PM          Subject:     RE: [nv-l] (NV
7.1.2) Bandwidth
                                                Utilization Collects





It looks like you have got a wrong expression.
It has to be like this.

OctetsIn (counter)* 8
-------------------------- * 100 = InUtilization
Speed of the interface
for 100Megs is (100 000 000)


This can't be more than 100 %


Dejan Mijailovic
Inquent Technologies



-----Original Message-----
From: David_Moon AT lincolnelectric DOT com
[mailto:David_Moon AT lincolnelectric DOT com]
Sent: Thursday, September 26, 2002 3:05 PM
To: nv-l AT lists.tivoli DOT com
Subject: [nv-l] (NV 7.1.2) Bandwidth Utilization Collects


List,

I am using the Bandwidth Utilization In and Out expressions in the SNMP
Data Collect section of Netview 7.1.2. I have set this up to collect data
on my routers and switches. However, the results I am seeing are totally
confusing me. I show a total utilization on some switches of over 100%. If
I am correct in my assumption, the value returned during the BW Utilization
SNMP collect should show the % of total utilization. I am trying to figure
out how the utilization on some switches is beyond their physical
capabilities. Either my network regularly eats it's wheaties or I have done
something wrong! Does anybody have any input on this. I have had a running
discussion about this with Netview support but they seem to think that this
problem has to do with the way our Cisco devices are reporting their IF
In/Out Octets. Our Network administrator swears that everything is correct
on the Cisco side.

Here is one of my events from the data collection.

Thu Sep 26 14:51:42 2002 LINCOLN_B.lecis D BandwidthUtilOut 41 threshold
exceeded (>300.00): 412.67 BandwidthUtilOut

All ideas are welcome!

Thanks,
Dave Moon
Systems Architect
Lincoln Electric Company


**************************************************************
 Note:  The information contained in this message may be
 privileged and confidential and protected from disclosure.
 If the reader of this message is not the intended recipient,
 or an employee or agent responsible for delivering this
 message to the intended recipient, you are hereby notified
 that any dissemination, distribution or copying of this
 communication is strictly prohibited. If you have received
 this communication in error, please notify us immediately by
 replying to the message and deleting it from your computer.
 Thank you,
 The Lincoln Electric Company
**************************************************************



---------------------------------------------------------------------
To unsubscribe, e-mail: nv-l-unsubscribe AT lists.tivoli DOT com
For additional commands, e-mail: nv-l-help AT lists.tivoli DOT com

*NOTE*
This is not an Offical Tivoli Support forum. If you need immediate
assistance from Tivoli please call the IBM Tivoli Software Group
help line at 1-800-TIVOLI8(848-6548)







---------------------------------------------------------------------
To unsubscribe, e-mail: nv-l-unsubscribe AT lists.tivoli DOT com
For additional commands, e-mail: nv-l-help AT lists.tivoli DOT com

*NOTE*
This is not an Offical Tivoli Support forum. If you need immediate
assistance from Tivoli please call the IBM Tivoli Software Group
help line at 1-800-TIVOLI8(848-6548)



---------------------------------------------------------------------
To unsubscribe, e-mail: nv-l-unsubscribe AT lists.tivoli DOT com
For additional commands, e-mail: nv-l-help AT lists.tivoli DOT com

*NOTE*
This is not an Offical Tivoli Support forum. If you need immediate
assistance from Tivoli please call the IBM Tivoli Software Group
help line at 1-800-TIVOLI8(848-6548)