Other folks do this sort of thing but the ruleset will not be what you
would use in UNIX. The pager node does not work on NT. Instead you have
to call the pager through a script (bat file) which uses the nvpage
command, and is run as an in-line action. It could also call nvmail. So
first you have to write your script or bat file. , and test it with inputs.
This by itself is not trivial. Then you have to decide what events you
want to correlate and design the logic of the ruleset, which is very
difficult if you have not seen one in UNIX. Then you have to get some
sympathetic soul (you could open a problem to Support and ask Level 2 to
help) to create a ruleset of the sort you want, and then ftp that to NT and
modify it to work. I know other folks have done this. Perhaps one of them
is reading.
What you are suggesting here -- not sending a page for interface down if
you got a router down -- would require more elaborate coding in UNIX even
than you suppose. There is no built-in template to match an interface with
the node it is from, so you would have to code up a script of some kind
even there to determine which interface is on which router. However, the
Interface down trap does contain the object id of its parent node, along
with other info in var 4, so you could run a script to obtain that object
id, then do an ovobjprint -o <object_id> and get the sending hostname that
way. As you can see from these comments, you would have to do some script
coding of your own even on UNIX, and even more on NT, so it might be just
as simple to write one big script to be run from the Interface down trap as
a windows or console app and forget the ruleset idea altogether. This
facility is not very mature in NT.
James Shanks
Team Leader, Level 3 Support
Tivoli NetView for UNIX and NT
jwittry@uswest.com on 10/05/2000 03:21:39 PM
To: James Shanks/Tivoli Systems@Tivoli Systems, lclark@us.ibm.com
cc:
Subject: Re: [NV-L] nvcoord ruleset and template documentation
Thanks for your response. It is greatly appreciated. I'm wanting to
correlate some critical events before sending pages or e-mails (i.e. don't
sent an alert message for every interface down event if I also get a router
down event for the same device in the same time frame). I can certainly do
the necessary scripting/programming to make this work but I didn't want to
re-invent the wheel if I didn't have to. Based on your comments however,
it
sounds like I have to do just that.
Thanks,
Jim Wittry
Qwest !nterprise Networking
(515) 241-3965
JWittry@USWEST.com
|