Ray you can do it using Mid Level Manager, but you have to be carreful with the
fact that your variable is an INTEGER.
In the MLM dialog box when you set your threshold collection you will have to
write "delta >" instead of ">"
(in fact "delta <condition>" instaed of "<condition>").
Banque Nationale de Paris
schafer AT tkg DOT com on 23/05/2000 18:02:50
Veuillez répondre à nv-l AT tkg DOT com
Pour : nv-l AT tkg DOT com
cc : (ccc : Alexis Henry/DOSI/BNP)
Objet : [NV-L] Integer Thresholds
L integrite de ce message n'etant pas assuree sur Internet, les societes du
groupe BNP Paribas ne peuvent etre tenues responsables de son contenu.
Si vous n'etes pas destinataire de ce message, merci de le detruire et
The integrity of this message cannot be guaranteed on the Internet, companies
of the BNP Paribas group can not therefore be considered responsible for the
If you are not the intended recipient of this message, then please delete it
and notify the sender.
Does anyone have any ideas how I can solve this problem:
I want to set an SNMP data collection threshold of 1000/hour for Cisco
routers bufferFails MIB variable. The variable is defined as an INTEGER
(not a COUNTER).
Here is what I tried so far, and why it doesn't work:
1. Use mib.coerce to force the variable to be a Counter. This doesn't
work because I want my threshold to be 1000/hour, snmpCollect
shows counters as units/second (can't find a way to change that!)
and Thresholds cannot be set for less than 1 (1000/hour is 0.278
2. Make an expression with the variable that is "coerced"into being a
counter by multiplying it by 3600. This doesn't work because for
MIB Expressions - NetView seems to ignore the mib.coerce file. It
just ends up multiplying the Integer value by 3600, not the delta
What I resorted to was collect the data as an Integer (default), use
snmpColDump to feed the data into an awk script that will calculate the
rate/hour of the integer change. However, because of buffered I/O, I
can't use the -f flag of snmpColDump to continually feed this data into
an awk script (it buffers up a bunch of lines before it sends it to awk
to process - meaning it only sends data every couple hours!).
Isn't there another way?
Ray Schafer | schafer AT tkg DOT com
The Kernel Group | Distributed Systems Management
NV-L List information (unsubscribing, policies, posting, digest version,
searchable archives): http://www.tkg.com/nv-l