Richard -

There is no good reason for netmon to hang servicing just 2500 objects.  There
are folks running well with 10 times that many nodes.  So I question whether the
problems you were seeing were truly the result of it having too much to do.  My
advice would be to call Support and make certain your have the latest fixes for
netmon and discuss your problems with a specialist.

As far as MLMs go, an MLM can monitor much more than it can discover (discovery
is limited to its won subnet but you can add many more objects to its collection
and it will poll those too) so the ideal places to put them is at the end of
slow links, so the polling traffic does not have to go over the link.   So they
still may be useful to you if it turns out that the bulk of your network is like
that.    But it sounds to me like you have some other more fundamental problem
with netmon and you should discuss that with Support if you can.  If you would
rather deploy a few MLMs first, that's your call.

James Shanks
Tivoli (NetView for UNIX) L3 Support



"Dearman, Richard" <rdearm1 AT UIC DOT EDU> on 08/10/99 07:46:22 PM

Please respond to Discussion of IBM NetView and POLYCENTER Manager on NetView
      <NV-L AT UCSBVM.UCSB DOT EDU>

To:   NV-L AT UCSBVM.UCSB DOT EDU
cc:    (bcc: James Shanks/Tivoli Systems)
Subject:  Re: Node status




1)  The polling interval was increased because 5 minutes was causing the
Netview server to get overloaded with ping requests and hang.

2)There are 2500 objects that Netview knows about.

3)There are over 30 subnets.  With the slowest connection being 10mb
ethernet.

Should the ping timeout values be changed I left those at the default?
Or am I going about this all wrong I figured putting more MLM's out would
allow me to distribute the load.

-----Original Message-----
From: James Shanks [mailto:James_Shanks AT TIVOLI DOT COM]
Sent: Tuesday, August 10, 1999 6:27 PM
To: NV-L AT UCSBVM.UCSB DOT EDU
Subject: Re: Node status


In order to answer this question you have to answer several prior ones.
(1) Why does it take you 15 minutes to know a node is down if netmon is
polling
every node at 5 minute intervals (the default)?  If you changed the polling
interval, why did you do it?
(2) How big is your network?  Do ovobjprint -S to find out how many objects
NetView knows about.
(3) People usually put MLMs where it makes sense to put them -- i.e on
remote
subnets grouped by location.  How is your network configured.  Many subnets?
A
few large ones?

James Shanks
Tivoli (NetView for UNIX) L3 Support



"Dearman, Richard" <rdearm1 AT UIC DOT EDU> on 08/10/99 07:21:33 PM

Please respond to Discussion of IBM NetView and POLYCENTER Manager on
NetView
      <NV-L AT UCSBVM.UCSB DOT EDU>

To:   NV-L AT UCSBVM.UCSB DOT EDU
cc:    (bcc: James Shanks/Tivoli Systems)
Subject:  Node status




Currently it takes me 15 minutes before I know a node is down.  I want to
decrease this time to less than 5 minutes.  I plan on deploying MLM's to
reduce the load on my Netview server but I don't know how many to deploy to
get the status level I desire of less than 5 minutes notice.  Is there any
document, chart or rule I can use to esitmate how many MLM's to deploy to
get a desired response time.

Thanks
Brian