Sean -
Take it easy. Your postings indicate that you have some fundamental
misunderstanding about all this and it is going to take time and effort to get
over that. I will see if I can help.
First, you are trying to guess what traps will be generated when a node or
interface goes down. I don't think Sun will send you any. Why should it? What
agent do you have in your network which is going to send the trap? If you have
a box, call it SunA and SunA keels over, what happens? I think the only agent
who is going to tell you about that is NetView (more specifically netmon), which
is why you purchased and implemented NetView in the first place. So unless you
have a Sun agent somewhere which is monitoring your network and sending in traps
when something goes down, you should forget about traps under the Sun Enterprise
Id. Only a Sun agent will send traps under the Sun Enterprise Id. And snmpdx
on Solaris does not do that, though NetView does. And when NetView sends a
trap, he will use only his enterprise Id. So your ruleset must reference
NetView traps. Does this make sense?
Second, if we are talking about NetView traps, then you can ignore the MIB
loader and mib2trap. These have no bearing on NetView traps. They can be used
to help interpret traps from other agents but that is all. And they play no
role whatsoever in a ruleset. The correlation daemon, nvcorrd, who runs your
ruleset, neither knows nor cares about what MIBs you have loaded, and except for
the fact that in an Event Attribute node you can access the trap severity and
category, he doesn't care how the trap is formatted for display in trapd.log or
the events window. All the ruleset cares about is the actual contents of the
trap, and what ruleset nodes you have selected to execute because of them.
So loading the Sun MIB will have no effect on your ruleset performance, and in
fact is irrelevant to what you are trying to do.
Third, in order to test your ruleset you must have it running somewhere. How
is yours running? Did you put into ESE.automation and ovstop/ovstart actionsvr?
Or did you create a dynamic workspace from an events window and specify your new
ruleset to run there? Those are the only two ways.
However you do it, you might want to turn on the nvcorrd trace -- issue the
command "nvcdebug -d all" -- so that you can look in the nvcorrd.alog and see
what happens when your trap is issued. And you could also look in the
nvaction.alog to see whether you get any errors or messages when the email
command is issued. It will take some detective work on your part to determine
where this is failing.
Fourth, what else are you checking for in your ruleset besides enterprise id and
specific trap number in a Trap Settings node? Are you looking for a particular
hostname too? Or what? Is your email action in a script or just a command? Is
the path to that script or command fully-qualified when you specify it?
I'm sorry that I cannot take problems to work on off this list. If you find
what I have told you truly confusing, then perhaps you should consider making a
call to Tivoli Support where people will look a your ruleset and those logs I
mentioned and help you get this implemented.
Good luck.
James Shanks
Tivoli (NetView for UNIX) L3 Support
Sean Aaron <sean.aaron@UCOP.EDU> on 09/03/99 01:31:44 PM
Please respond to Discussion of IBM NetView and POLYCENTER Manager on NetView
<NV-L@UCSBVM.UCSB.EDU>
To: NV-L@UCSBVM.UCSB.EDU
cc: (bcc: James Shanks/Tivoli Systems)
Subject: Re: Reality check requested
James Shanks wrote:
>
> No they are not equivalent. The traps you sent use the NetView Enterprise Id
--
> Interface Down and Node Down are NetView traps. That's what your ruleset
should
> being using. You are not going to get any generic enterprise traps from
> NetView, or from Sun. They will use their actual Enterprise Id.
>
> James Shanks
> Tivoli (NetView for UNIX) L3 Support
>
I don't understand. Are you saying I should be using the netView6000
traps for node down or link down, or are you saying that I should create
traps specifically for the Sun oid that matches the host in question?
I tried using the netView6000 traps, but when I try testing with the
traps listed in "diagnose->send event to trap daemon" I still get no
e-mail to my pager. If it's a matter of having a matching
enterprise/trap collection, why wouldn't my loading the sun.mib and
snmpdx.mib add the traps? Do I need to use mib2trap to create a Sun
enterprise entry?
--
Sean Aaron
UNIX System Administrator
University of California
Office of the President
|