Thanks, Christian. That should work with the mib expression,
I think. The MIB Datacollector only thresholds in a positive
direction, and you want to threshold in a negative direction.
So in /usr/OV/conf/mibExpr.conf, you could add this entry:
UpTimeDown \
" the negation of sysUptime" \
.1.3.6.1.2.1.1.3. -1 *
The first record is a label, the second is a description,
and the third is a formula in post-fix notation. (I think
this is what they called 'reverse Polish' in 9th grade math).
You have to stop/start snmpCollect to pick this up. Check
snmpCol.trace for syntax errors. Now in the data collection
and thresholding dialog, Add a Collection but chose expression
rather than MIB variable, and it should offer you that label.
Collect on the label for the expression rather than the MIB
variable directly. Now the value when things are going well
will be -100, and higher than that (but less than 0) when
there has been an outage. Threshold on greater than -100.
There is an outside chance that the mib.coerce setting is not
honored when using expressions, but the last time I looked
deeply into this I believe I came to the conclusion that it
was working correctly. Let us know.
Cordially,
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
Hi!
In mib.coerce i have added this
.1.3.6.1.2.1.1.3 COUNTER
When doing this netview makes a calculation on the value it get's from
the mib-variable. The result is the rate of change over that period.
-- man mib.coerce --
For example, if the sampled counter value is 1000 at 12:00:00 and
1600 at 12:00:30, the rate of change over that period is
(1600-1000) / 30 seconds, or 20 per second. The data collector
stores the value 20.
This counter calculation yields the rate of change per second.
The rate of change is stored instead of the raw counter values
because it is more important to know how fast counters are
changing than what their value is at a given time. If the value
at the second sample is smaller than the value at the first
sample, the first sample is ignored because the counter has
wrapped or because the node has been re-initialized.
-- man mib.coerce --
So if I collect sysUptime every minut I get a value of 6000 / 60 = 100.
But if the unit has been down say 20 seconds i get 4000 / 60 = ~67
Could you clarify the mibExpr.conf tihng some more, maybe with a little
example =)
/Christian
-----Original Message-----
From: Leslie Clark [mailto:lclark@US.IBM.COM]
Sent: Wednesday, September 29, 1999 1:49 PM
To: NV-L@UCSBVM.UCSB.EDU
Subject: Re: Data Collections
This sounds promising, Christian. But you have to tell us what you did in
mib.coerce. At least I cannot imagine what you mean.
But usually when I want to turn around the values returned by data
collection
I do it with an expression in mibExpr.conf. So maybe I get my value x -1
so the thresholds work in the right direction. But I don't know if that
would
work
in your case, since I don't understand what you have done with mib.coerce.
Cordially,
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
Hi!
In my neverending search for a way to calculate avaliability, I need some
help =)
I have added sysUptime in the mib.coerce file so all polls that are made and
the unit has
been up all the time since the last poll i get the value 100, if the unit
has been rebooted or
anyhting else that resets the sysUptime has ocurred, the value is not logged
in my data-collection.
I would like to have it the other way around, when it's 100 it's discarded
and when it's something
else it's logged. Then it shouldn't be to hard to make something that dumps
the data collection
and calculates the availabilty.
Is this possible to do, or am I forced to make a huge list of all
successfull polls and then make a script
that finds the glitches?
-----------------------------------------
Christian Frantsen
Technical Operations
Internoc Scandinavia AB
Tel: +46-36-194843
Fax: +46-36-194651
http://www.internoc.se
|