nv-l
[Top] [All Lists]

Re: problems and questions with NV6.0.1

To: nv-l@lists.tivoli.com
Subject: Re: problems and questions with NV6.0.1
From: "Leslie Clark" <lclark@us.ibm.com>
Date: Mon, 27 Nov 2000 07:05:22 -0500
See imbedded comments

Cordially,

Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
Detroit


"Stefan.Beck" <STEFAN.BECK@Allianz.de>@tkg.com on 11/27/2000 02:41:42 AM

Please respond to IBM NetView Discussion <nv-l@tkg.com>

Sent by:  owner-nv-l@tkg.com


To:   " - *nv-l@tkg.com" <nv-l@tkg.com>
cc:
Subject:  [NV-L] problems and questions with NV6.0.1



Hello NetViewers,

we are evaluating NV6.0.1 at a customer site on AIX4.3.3 on a SP2. We are
using
the new libnvsnmp.a .

We have the following problems:
1) netmon won't discover addresses from the seedfile that are marked with a
'$'
for snmp-status-polling.
Everything works perfectly, if we start without the '$'. Inserting the '$'
__after__ netmon has discovered the node,
netmon does status monitoring via snmp (works great).

>> That is working as designed. The $ is a management instruction, not an
>> instruction for discovery. To force the discovery of the address, put it
>> in twice, once with the $ for snmp management, once without for
discovery.

2) netmon won't discover loopback interfaces with an ip address starting
with
127.x.x.x (but not equal 127.0.0.1).
According to the release notes from V6.0.1 (page 10) I thought that netmon
should discover those interfaces.
(Yes, I know you can't ping them (no route) and most likely there's no use
in
having an icon for that interfaces in the map,
but netmon does not work that way that is described in the manual)

>> I've seen that, too. I suppose it assumes those are traditional
loopbacks,
>> which it intentionally does not draw. Otherwise it would be drawing it
>> and calling it Down. The most you could expect from opening a ticket on
>> this is a documentation APAR.

3) When we start the GUI, we get error messages from the maptreeserver:
Starting maptreeserver
signal Handler set
OvwEventDispatcher::EnableJavaCallbacks: socket()(error:12)
OvwEventDispatcher::EnableJavaCallbacks: bind():12
Parsing configuration file
/usr/OV/conf/C/explore.confOvwInterface::Connect: con
nect() (error: 22)
OvwEventDispatcher::EnableJavaCallbacks:  accept() : 22
Server is ready to accept connections on port 62535

I haven't investigated in that issue any further since the Java-GUI seems
to
work fine.
Is anybody  receiving the same messages (in the nv6000.log)

>> Yes, those are the normal noises for 6.01

4) We have had a problem with the location.conf:
We started discovery from the scratch with a seedfile and a location.conf.
Everything worked fine (no error messages in the location.log) and every
network was placed in
apropriate location. After a few days we recognized (via an ovtopofix) that
several interfaces were not
showing up in the map. We saw (in the nett.log0x) that ovw couldn't draw
the
segments and networks for the missing
interface icons and at that moment there were error messages in the
location.log that some locations already exist.
After removing the location.conf, __all__ missing icons showed up.

Has anybody had the same problem?

>> I saw this at 6.0 once, and have been very cautious in my use of
>> location.conf. If you do a locate on those missing interfaces, you
>> may find reference to them in submaps that don't exist, or are
>> orphaned (no parent). When you open those submaps, you can correct
>> the parentage with Edit..Modify/Describe..Submap. Check to see
>> that all of the locations you defined actually showed up. I believe
>> this problem can be caused by subtle errors in defining networks in
>> the location.conf. If you have a good example, you should take it up
>> with Support. My favorite way to use location.conf is just to generate
>> the heirarchy of location icons, with one specific network in each one,
>> one that will cause a router to be attached to it. And after that,
>> remove the location.conf.

5) Question: If a node is marked with a '$' for snmp status polling in the
seedfile and the observation for that node is delegated
to an MLM, what will happen?

>> It will revert to pinging, since MLM does not participate in the
>> snmp polling enhancement.

Any help and suggestions are welcome.

regards
Stefan Beck
_________________________________________________________________________
NV-L List information and Archives: http://www.tkg.com/nv-l


<Prev in Thread] Current Thread [Next in Thread>

Archive operated by Skills 1st Ltd

See also: The NetView Web