Re: [nv-l] Need guidance in troubleshooting the reception of traps

[James Shanks <jshanks@us.ibm.com>]

There's not much to the flow if you
are just interested in whether an event made it to the trapd.log or the
event database.  trapd owns both on Windows, so the flow is MS Trap
Service to trapd and that's it.  But before you start tracing, go
to the nv.log and see whether trapd recorded any errors.  If there
are none, then you have to trace.

 To see whether trapd got a trap,
and what he got, you have to get a trapd.trace.  The only way to do
that on Windows is to stop trapd, edit the ovsuf file, and add 
        ,-x,-t\usr\ov\log\trapd.trace
after the other options you see there.
  The when you restart trapd, he'll keep a trace record of what he
gets and you'll see the trap contents there if he gets it.   We expanded
that trace just recently, so I hope you have at least 7.1.4 FixPack1 installed.

There is no troubleshooting you can
do of the MS Trap Service externally  (Microsoft doesn't believe that
their code is ever in error apparently).  All you can do there is
get an external trace of what's coming in to the box, using some other
tool.  My recommendation is for the ethereal analyzer which you can
download free from the web (http://www.ethereal.com).  It's a simple
install and the only filter you need create to look at incoming traps is
just "udp and port 162".   That's how we've solve problems
with weird stuff coming from the Trap Service before.  If you don't
see your trap in the ethereal trace, then it never got there.

HTH

James Shanks
Level 3 Support  for Tivoli NetView for UNIX and Windows
Tivoli Software / IBM Software Group

"Clays, Michael" <Clays.Michael@jdirving.com>
Sent by: owner-nv-l@lists.us.ibm.com

09/02/2004 07:01 PM

Please respond to
nv-l

To	nv-l@lists.us.ibm.com
cc
Subject	[nv-l] Need guidance in troubleshooting the reception of traps

I am looking for some guidance in troubleshooting
the reception of traps on our Windows 2000 SP4 Netview 7.1.4 box. I am
not sure at which point the trap gets lost.
 
I have a server that I am using to send the
traps from, I have it pointed to both a lab and a production Netview Windows
server. The traps work fine on the lab Netview, I see them in trapd.log
every time. The production side is where my problems lie. I do not see
any indication in trapd.log that the trap was ever received. 
 
If, on the production Netview server, I restart
the Windows SNMP Trap Service, and Netview, then send a trap I will see
it in trapd.log. But within a short time any further traps I send do not
show up. I do not know if the problem is with the Windows Trap service
or in Netview.
 
What I am looking for is an overview of the
flow the trap takes when it arrives at the server and how to troubleshoot
each step to see where the failure is occurring.
 
Any assistance or pointers to troubleshooting
documentation would be greatly appreciated.