Nagios Support Forum

I have the same problem on my production network.

I solved the problem using some perl script:

1) one run the nagios check(s) and store the result in a local buffer;

2) one open a connection from the nagios central server to the remote
server to monitor (the appliance, in your case) via ssh, and retrieve
all the results, storing them in a local buffer (on the monitor server);

3) one other, the last, scan the local buffer on the monitor server and
submit the passive results to nagios, via pipe file.

I am fully satisfied of this solution:

.) I do not loose any result, also if it arrives very late (days,
weeks, ...);

.) I keep track of all the monitoring jobs, with the correct date/time
(are on the log message);

.) I have a warning if I do not receive a test for more then half an hour,
but this is optional;

.) We decided to have the connection started from the nagios server to the
remote server to test because of security policy (the nagios server is
behind a firewall), but this can be easily changed.

Are you interested in these scripts? If yes I shouldn't have problems
to send them, to you or to the list if someone else is interested.

They are a bit more complicated that the needs, because they were
originally made for another, more complicated, job, but they are
fully usable and modifiable. Let me know!

Bye
Stefano

On Sat, Mar 08, 2003 at 12:08:19PM -0800, Thomas wrote:

> From: Thomas
> To: [email protected]
> Date: 08 Mar 2003 20:20:06 +0100
> Subject: [Nagios-devel] Nsca, passive service checks
>
> Hi all,
>
> I already use Nagios to monitor many hosts and many services form the
> office of my company. I would like to use a kind of
> appliance to monitor distant networks. I think the best way to do it is
> to use the passive service checks : nagios would be running on the
> appliance and send service check results to the central monitoring
> server who will handle notifications.
>
> My problem is what would happen if a the connection between the central
> server and the appliance is down (for example there could be a problem
> of routing, or an IPSec VPN could fail) ? The nagios on the appliance
> will still be working and monitoring the hosts on its side of the
> network, and may be still connected to the public network but won't be
> abble to send the results to the central server.
>
> The most important for me is not notifications, I know that there are
> many ways to handle that, I could activate notification on the appliance
> for example. My problem is that I need to keep track of all the
> monitoring job that is done during the network breakdown. I need the
> central server to know all what happened on the distant network during
> breakdown once everything is repaired, in order to be abble to generate
> reports. I can't have "holes" in my reports for the hosts and services
> on the distant network.
>
> Maybe I could program a kind of buffer that would send all service check
> results after the breakdown ? Is this a good solution ? But I think that
> nsca has a "MAX PACKET AGE" option that can't exceed 900 seconds. Maybe
> it is possible to modify that at compile time ? Is there any other
> solution ? What about copy/paste of logs ?
>
> I have read the Nagios documentation very carefully but maybe I missed
> something.
>
> Thanks for your replies !
>
> Thomas





This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]