Re: Netmon exited on Signal 4

1999-12-13 09:31:35
Subject: Re: Netmon exited on Signal 4
From: Leslie Clark <lclark AT US.IBM DOT COM>
To: nv-l AT lists.tivoli DOT com
Date: Mon, 13 Dec 1999 09:31:35 -0500
There is a new-ish problem that causes netmon to die with a signal 4
that you can perhaps avoid until a fix becomes available. I have seen it
at 5.1.2 but it may have started at 5.1.1.

Apparently there are circumstances under which a config poll of a node
with both unnumbered serial interafaces AND hsrp interfaces will confuse
netmon and it will die with a signal 4.  It is APAR IY04540.

Since you cannot disable hsrp support, the alternative is to disable
unnumbered support in netmon's configuration. I'm not sure what happens
if you turn it on to get your discovery done and then turn it off.

Follow up with Support to verify that this is your problem and check status
of an e-fix.


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

A signal 4 is an illegal instruction (see the man page on signal or kill),
probably means that there was a memory overlay somewhere -- either netmon
overlaid his own memory or someone else did it for him -- so that when the
instruction which was at that location came to be executed, it was
Please don't confuse a sig 4 with a compile error -- the instruction was
at the time the code was compiled.  It became illegal because all or part
of it
was changed or replaced at run-time.

What can you do about it?  Well, very little.   You should probably call
as the first step, because the situation is basically like this:

 If you are not at 5.1.2, then you should get there, because (a) the fix
already be in that level and (b) even if it is not, any fix from this point
will require that level as a base.

 If you are at 5.1.2, then your only option is to call Support and gather
data so they can isolate the problem.  Look in /usr/OV/PD/cores/netmon.  Do
have a core?  If so, that will help.  If you are on AIX, you can run
/usr/OV/service/readcore against the core and get a report which gives the
stack so Support can tell you if it has been reported before.  But readcore
requires the AIX utility, dbx, be installed.  You get that off the AIX CD
if you
haven't installed it already -- its part of the application debugger set.

If you don't have AIX, but Solaris or Digital Unix, you must call Support
that they can give you utilities to help with the same job.

And you could also format the nettl logs:
     netfmt -f /usr/OV/log/nettl.LOG00 > "myfile1"
     netfmt -f /usr/OV/log/nettl.LOG01 > "myfile2"
and look for errors about the time that this happened.  You may also be
asked to
run the netmon trace with all options until this happens again so that
can see what leads up to it.

Of course, you could do all this even if you were not at 5.1.2, but it may
pointless, if Support tells you they already have a fix and that's where it

James Shanks
Tivoli (NetView for UNIX) L3 Support

Scott Hammons <SHammons AT MCBRIDE DOT COM> on 12/10/99 03:52:10 PM

Please respond to Discussion of IBM NetView and POLYCENTER Manager on

cc:    (bcc: James Shanks/Tivoli Systems)
Subject:  Netmon exited on Signal 4

This has happened twice in the last week.  Has anyone seen this error
before.  I checked the Diagnosis Guide and it has the reason for exited on
signal 11, but not exited on signal 4.

Scott Hammons
McBride and Associates, Inc.

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