[Top] [All Lists]

Re: Maximum number of events?

To: nv-l@lists.tivoli.com
Subject: Re: Maximum number of events?
From: James_Shanks@tivoli.com
Date: Mon, 2 Apr 2001 15:35:01 -0400
Trust me, Les.  I know what this message means.  netmon does not know how
to stop sending traps.   trapd does get raw traps from netmon, and other
sources, and he forwards some or all of those to other "interested
parties".  ipmap for example gets all netmon traps.  And netmon used to get
(in 5.1) lots of things and just throw them away.  Now he gets Link Up and
Link Down traps from other agents.  He uses these to immediately go check
those nodes so as to respond to changes in your network quicker -- you
don't have to wait for netmon's polling cycle to come around to those guys
again. That is a feature of 6.0.2.  But netmon cannot process these traps
while he is initializing.  So if you get a whole bunch of them while he is
still doing his config checks, then they fill up netmon's internal queue in
trapd and trapd forces him off to protect himself.  Which is just as well,
because netmon will find out about the current state of those  interfaces
during his init routines anyway.

If you like, you can raise the size of netmon's queue (and everybody else's
too -- it's a global thing) by adjusting the application queue buffer size
in trapd, but it doesn't pay to make it real big anymore.  The default size
is 2000.  You can make it huge, say 20,000 if you like, but that means that
the first thing netmon will do after he initializes is process all those
traps.  May mean a lot of pointless cycles.  But if you think that netmon
should be just about ready to go when he gets forced off and making the
buffer bigger, say double or triple, will allow him to stay connected and
catch up, then give it a shot.

If you want, you can use trapd.trace (just type in "trapd -T" from the
command line as root to toggle it on and off) to see what other
applications have connected to trapd.  Each time a trap is queued to a
particular appl, you see a message which says that the queue for that appl
(identified by an internal number and a PID) is now such and such big.
Using "ps -ef" you can see what that PID is.  It can be enlightening.

James Shanks
Team Leader, Level 3 Support
 Tivoli NetView for UNIX and NT

"Les Dickert" <lesdickert@hotmail.com>@tkg.com on 04/02/2001 02:20:41 PM

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

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

To:   nv-l@tkg.com
Subject:  Re: [NV-L] Maximum number of events?


I thought netmon sent traps to trapd, not
vice versa.

Could the error message mean that trapd is
not processing the traps from netmon (probably
due to an external trap storm) and that netmon
is going to quit sending them for a while?

Les Dickert
Verisign Consulting

>From: James_Shanks@tivoli.com
>Reply-To: IBM NetView Discussion <nv-l@tkg.com>
>To: IBM NetView Discussion <nv-l@tkg.com>
>Subject: Re: [NV-L] Maximum number of events?
>Date: Mon, 2 Apr 2001 10:38:55 -0400
>That depends on what is going on.  The message means that netmon is not
>processing the traps that trapd is sending to him, probably because he is
>busy doing something else.  When his queue gets full, trapd terminates his
>connection, so that he himself does not wind up in an out-of-memory
>condition.  If you have just restarted netmon then this may be entirely
>normal, because he has to do all of his initialization and config checks
>before he can process traps.  When he is done with his init stuff, he will
>re-connect to trapd and start paying attention to the incoming trap flow.
>If you have not just restarted netmon, then you may have a trap storm
>problem or a netmon hang of some kind to investigate.  I would start with
>the netmon trace and see what netmon is doing in that case, and call
>Support if you need help.
>James Shanks
>Team Leader, Level 3 Support
>  Tivoli NetView for UNIX and NT
>"Treptow, Craig" <Treptow.Craig@principal.com>@tkg.com on 04/02/2001
>10:04:51 AM
>Please respond to IBM NetView Discussion <nv-l@tkg.com>
>Sent by:  owner-nv-l@tkg.com
>To:   "NetView List (E-mail)" <nv-l@tkg.com>
>Subject:  [NV-L] Maximum number of events?
>Hi.  We are running Netview 6.0.2 on AIX 4.3.3.  I just saw an event
>by with this message:
>"netmon reached maximum number of outstanding events, disconnecting from
>The processes are still running, and things still appear ok.  Should I be
>Craig Treptow
>Principal Financial Group
>I/S Network Administration
>(515) 247-6207
>NV-L List information and Archives: http://www.tkg.com/nv-l

Get your FREE download of MSN Explorer at http://explorer.msn.com

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