IF YOU ARE ON UNIX, then this is what the APM is for. If you are on NT
you do not have the APM.
On NT, to do this by hand, here is what you have to do:
1) set up data collection to monitor the mib variable, and send one trap
for
the threshold and one trap for the rearm. Trap numbers 1000 - 1999 are
reserved for your use. They go in odd-even pairs.
2) add an icon (eg a 'Software' icon) to the node-level submap, next to
the interface cards. Note its Selection Name.
3) Add the traps you referenced in step 1 to the Netview enterprise,
generic 6, specific xxxx. Give them an automatic action. That action should
probably be to execute a script or bat file that issues the snmptrap
command. The trap that it issues would be the one that sets status,
generic 6 specific 58916871. See the online help for details. You need
to be able to arrive at the name of the application object from the name
of the
node, so take that into consideration when creating the object.
Since the status source for the nodes is 'compound propagate',
the status of your icon will contribute to the overall status of the node.
On unix, you have to install and configure MLM, then configure the
C5d daemon to start. Then you can use the APM function to select
the mib variable to monitor, and the collection of target nodes to
which that policy applies, and distribute it to the MLM that watches
those boxes. This automatically adds the icons and the traps are
handled for you. It also adds a grouping to the top level map to
show status just from the point of view of your thresholds.
Cordially,
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
Detroit
========================================================
Thanks for responding, Jim:
When I run NetView using the demo copy I received from Tivoli I create a
submap for selected nodes
of our lab. I can see their "pinged" status OK and when I click on them I
see the icon for their respective
network connection. What I;d really like is another icon that collects
some bit of data from each Alpha-workstation's MIB periodically. This
would simulate status behavior if we had written a more complex
application-specific
agent.
You see I'm trying to just demo capability and don;t need to have the
details worked out. Raytheon presently has a
box called the Monitoring and Control Display that is a proprietary
"NetView-like" manager of its application s/w
runnning on the many different types of servers that make up an Air Traffic
Control (ATC) System. We'd like to investigate replacing, or augmenting it
with application monitoring using NetView.
I realize I can use the MIB Browser and monitor a MIB variable on a machine
but that is different then getting status from a number of machines over
time. I'm sorry about my ignorance in some of these matters. All I have
is the demo
software and whatever training I could give myself in these matters. What
we have going on here is a research lab where we are investigating new
technologies we can use for our ATC Systems. Tivoli's Federal Group has
given us your software to try. So far most of my efforts have been to
alter the GUI to our rquirements, now I'm moving into other areas. Perhaps
I've missed something.
How can NetView be setup to use snmpCollect to set the status of an icon
for each Alpha workstation's agent that would depict some form of the
machine's status?
Many thanks for your response
Bob
James Shanks <James_Shanks@TIVOLI.COM> on 01/29/2000 02:36:12 PM
Please respond to Discussion of IBM NetView and POLYCENTER Manager on
NetView <NV-L@UCSBVM.UCSB.EDU>
To: NV-L@UCSBVM.UCSB.EDU
cc: (bcc: Robert Stamm/RES/Raytheon/US)
Subject: Re: Simple agent
I don't understand why you require sample source code to do this. In fact,
just
about everything you've said here baffles me. If these DEC/Compaq UNIX
boxes
run Digital UNIX or Tru64, then there is a standard SNMP agent which can be
installed on them, and usually is by default. It supports all the usual
MIB2
stuff that everyone else supports. If this is running, then you can use
NetView's own snmpCollect to gather MIB data on a periodic basis and do
with it
what you will, without putting any new agent on the box . You could also
install MLM on one of them and collect data that way, if being closer to
the
source is the issue. So you don't need any sample code at all.
To put it another way , you are asking for a sample agent to do what
NetView
already does, but that doesn't make sense to me. So just what is it
you are
really trying to do here? Collect data from a private MIB of your own
making?
Are you asking for sample code for an agent which will populate your
private MIB
and respond when some management station (like NetView) makes an SNMP
query for
that data? Is that why you need source code?
James Shanks
Tivoli (NetView for UNIX) L3 Support
Bob Stamm <Robert_Stamm@RES.RAYTHEON.COM> on 01/28/2000 01:22:45 PM
Please respond to Discussion of IBM NetView and POLYCENTER Manager on
NetView
<NV-L@UCSBVM.UCSB.EDU>
To: NV-L@UCSBVM.UCSB.EDU
cc: (bcc: James Shanks/Tivoli Systems)
Subject: Simple agent
I am an engineer for Raytheon Co. Here we build Air Traffic Control
Systems and I have
installed NetView for a demo of how we might use it for Control and
Monitoring of the DEC
Alpha Unix boxes (now Compaq Computers) we use to power our ATC Controller
consoles.
What I wonder is if anyone has a sample simple agent (written in C) that
can be used by me
to query stuff in the MIB and periodically report to NetView a status of
some MIB variable.
Upon receipt of this event I'd like to change a symbol representing the
status of that agent.
I'd probably only run the agent on a single workstation since I have to put
it into an operational
ATC installation in Munich and I'm sure I'll have limited access to their
equipment.
Is there a source of information (sample code) that I can refer to to help
me put together a
quick demo? The only resources I've got are a demo copy of NetView atnd my
own wits ...
A piece of working samplle code would go a long way to help me pull this
off.
Bob
|