You have to install the MLM. The code is on the Netview CD in a
separate directory. Copy the image to disk and use smit to install it.
There is an upgraded one on the 5.1.1 CD. Also there is a patched
version of the binary available from support as on e-fix to correct some
problem. I would get that too.
Starting up APM will set the trap destination for you. Make sure the
server has snmp write capability to wherever the MLM is installed.
I usually find that I have to add a entry to /etc/snmpd.conf to allow
write not only from 127.0.0.1 but also from the regular IP address
of the server even if it is on the server. (MLM's trap destination should
be the Netview server).
The MLM can be installed anywhere. And you need to use the Collection
Editor to make sure that the MLM is responsible for the devices you want
to monitor. I usually make a collection of those devices, then modify the
mlm domain collection to include that device collection. When you distribute
a policy, you direct it to the Collection of devices to be monitored, then
APM translates that into the name of the MLM(s) responsible for that collection,
based on the mlm domain collections.
Note that this is a different job for the MLM than the one usually being
discussed
which is the offload of polling. So there is no need to tell netmon about it at
all.
Cordially,
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
The method of using APM to threshold MIB variables as mentioned below
interests me.
A couple of questions though. If I want to threshold standard MIB-II or
private MIB variables
(not SIA/MLM MIBs) in routers through APM, do I need to have MLM installed
on the same
box as Netview?
Anything else I need to install/setup (apart from registering C5d daemon)?
Environment: AIX 4.2.1 NetView 5.1.1
Thanks!
Regards,
Haakan
Leslie Clark <lclark@US.IBM.COM> den 99-08-20 03:49 GMT
Sänd svar till Discussion of IBM NetView and POLYCENTER Manager on NetView
<NV-L@UCSBVM.UCSB.EDU>
Till: NV-L@UCSBVM.UCSB.EDU
Kopia: (blank: Håkan Andersson/P00HAA/HK_IT/STOCKHOLM/ICA/SE)
Netview provides a couple of ways to make noise when a mib variable
exceeds a threshold. How are you monitoring the variable? If you are
using Data Collection & Thresholding (snmpCollect), then there is a
threshold trap and a rearm trap that can be automatically generated
as specified in the snmpCollect dialog. You can also specify your own
trap in that dialog (start with 1001/1002) in the netview6000 enterprise.
snmpCollect generates the events.
My favorite way, if you also want the icon to change color when the
threshold is exeeded, is to use the Agent Policy Manager (C5d daemon)
that works with MLM. It is under Tools. It is more complicated to set up,
but very powerful. It adds an icon to the node submap where it contributes
to the node status the same way the interface cards do. You specify the
threshold rearm traps that are under the MLM-Threshold enterprise, and
they automatically change the status of the icons.
All of these events show up in the events display, and can be customized
under Options...Event Customization or dealt with in rulesets.
Read more about it in the Admin Guide.
A couple of good Redbooks:
__ Redbooks SG24-4515 Examples of Using Netview for AIX V4 and SG24-4398
IBM Systems Monitor: Anatomy of a Smart Agent are available from IBM
and still relevant. See http://www.redbooks.ibm.com
__ Redbook SG24-5285 Integrated Management Solutions Using Netview V5.1
came out 3/99. See http://www.redbooks.ibm.com
Cordially,
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
I am confused about what and how I should use events to send an alert when
CRCs
are building in a router. When creating the event should I use
IBM_NVICE_EV, or
do I create a new one using LocIfInCRC MIB associated with the router? Can
I
set this up network wide or does each individual router have to be setup
individually? Or should I use a different tool all together? Thanks in
advance.
|