Re: Send a trap on non failure
David, We all read this list. > Put another way, if I could call remote monitor and tell it precisely which > server/hostgroup/service to check, and just have it report based on the > status of that service, it would be closer to useful. Closer to hard MONconfig synchronisation. It is why remote.monitor does whole status. Use mon trap feature for doing what you want and go to "hard" MONconfig synchronisation :-( man mon: "Traps are handled similarly to monitors: a trap sends an operational status, summary line, and description text, and mon generates an alert or upalert as necessary." -- Au revoir, 33 (0) 2 99 78 62 49 Gilles Lamiral. France, L'Hermitage (35590) 33 (0) 6 20 79 76 06
RE: Send a trap on non failure
--On Wednesday, June 26, 2002 12:39 PM +0200 Laurent Combe
<[EMAIL PROTECTED]> wrote:
> with remote.monitor (v1.7)
> you can use arg called (--bigsummary)
>
> with the following output :
>
># --bigsummary : flag to extend the summary of this monitor
>#return for each failed mon server the list of the
>#failed. Like : host1([g1:s1{sum}|s3{sum}][g4:s5{sum}])
> ...
>
> where g1,g2,... are hostgroups
> where s1,s2,... are services
> and sum the summary
>
> does it help ?
> do you want i mail you directly this release ?
>
I think you misunderstood what I meant, maybe I wasn't clear.
My understanding of remote.monitor is that the "master" server would only
end up with a single hostgroup/service entry, and all the failure
information about the sub-servers would be in the summary for that single
entry. I want the "master" to have entries for every hostgroup/service
monitored by the sub-servers. That way if a service monitored by
sub-server A is actually dependent on a service monitored by sub-server B,
the master server can verify that dependency before sending an alert. Also
a client (like mon.cgi) can talk to just the master server, and get back
full status information rather then just a success/failure status bit for
the entire system (plus a summary containing the actual details). (And the
full status information, i.e. the FULL output from the failed montiors,
would be unavailable I think.)
Put another way, if I could call remote monitor and tell it precisely which
server/hostgroup/service to check, and just have it report based on the
status of that service, it would be closer to useful. But I still think
we'd lose the exit codes from the original monitor scripts, but maybe there
is a way to get those. I haven't actually used remote.monitor, I've just
looked at what the (sparse) documentation says it does.
-David
David Nolan<*>[EMAIL PROTECTED]
curses: May you be forced to grep the termcap of an unclean yacc while
a herd of rogue emacs fsck your troff and vgrind your pathalias!
RE: Send a trap on non failure
> David Nolan wrote:
> That doesn't do what I want. As far as I can tell remote.monitor loses
all
> the hostgroup/service information. On the server running remote.monitor
> everything from the remote server would show up in the same
> watchgroup/service.
> I want the master server to inherit the complete information about the
> status of every hostgroup and service, so it can do its own dependency
> checking, and make its own decisions regarding what to alert about.
with remote.monitor (v1.7)
you can use arg called (--bigsummary)
with the following output :
# --bigsummary : flag to extend the summary of this monitor
#return for each failed mon server the list of the
#failed. Like : host1([g1:s1{sum}|s3{sum}][g4:s5{sum}])
...
where g1,g2,... are hostgroups
where s1,s2,... are services
and sum the summary
does it help ?
do you want i mail you directly this release ?
bye
laurent combe
(ps: bonjour gilles!)
Re: Send a trap on non failure
David Nolan wrote: > As part of the MON deployment project I'm working on we've actually been > talking about about wanting exactly this sort of thing for our > multiple-mon-servers infrastructure. Use the remote.monitor -- Au revoir, 33 (0) 2 99 78 62 49 Gilles Lamiral. France, L'Hermitage (35590) 33 (0) 6 20 79 76 06
Re: Send a trap on non failure
--On Monday, June 24, 2002 10:33 AM +0200 Daniel Luginger
<[EMAIL PROTECTED]> wrote:
> Hi ,
> i want to send a trap if the status of a service is ok.
>
> Is there a possibility to make an entry in the .cf file like this or in
> an other way:
>
> period wd {Sun-Sat}
> alert trap.alert // If the service is false
> alertok trap.alert // If the service is ok
>
>
As part of the MON deployment project I'm working on we've actually been
talking about about wanting exactly this sort of thing for our
multiple-mon-servers infrastructure. We're thinking about adding a new
configuration item, something like:
redistribute trap.alert server1.foo.com server2.foo.com
The idea is that every time the monitor script is called the result should
be propagated to the servers listed, regardless of time of day or status of
the service. (Lines of this form would belong OUTSIDE period definitions)
Anyone have any comments on that idea?
-David Nolan
Network Software Developer
Computing Services
Carnegie Mellon University
RE: Send a trap on non failure
What I have done :
On status ok of my monitor, I exit with 2 instead of 0.
In my mon.cf :
alert exit=2 trap.alert -o 1 my_mon_central_server
So I send a trap with status = ok and the informations ...
I use exit=2 because
alert exit=0 trap.alert -o 1 my_mon_central_server
does not seem to work.
-Message d'origine-
De : Daniel Luginger [mailto:[EMAIL PROTECTED]]
Envoyé : lundi 24 juin 2002 10:33
À : [EMAIL PROTECTED]
Objet : Send a trap on non failure
Hi ,
i want to send a trap if the status of a service is ok.
Is there a possibility to make an entry in the .cf file like this or in
an other way:
period wd {Sun-Sat}
alert trap.alert // If the service is false
alertok trap.alert // If the service is ok
Any idea is welcome.
Daniel
-
Ce message est confidentiel ; son contenu ne represente en aucun
cas un engagement de la part de Dexia Credit Local ou de Dexia
CLF Banque. Toute publication, utilisation ou diffusion, meme
partielle, doit etre autorisee prealablement par l'emetteur. Si
vous n'etes pas destinataire de ce message, merci d'en avertir
immediatement l'expediteur.
This message is confidential ; its contents do not constitute a
commitment by Dexia Credit Local or Dexia CLF Banque. Any
unauthorised disclosure, use or dissemination, either whole or
partial, is prohibited. If you are not the intended recipient of
the message, please notify the sender immediately.
-
Consultez notre site internet www.dexia-clf.fr
La cle d'une gestion optimisee du secteur public local
-
Send a trap on non failure
Hi ,
i want to send a trap if the status of a service is ok.
Is there a possibility to make an entry in the .cf file like this or in
an other way:
period wd {Sun-Sat}
alert trap.alert // If the service is false
alertok trap.alert // If the service is ok
Any idea is welcome.
Daniel
