[Nagiosplug-devel] Further development of nsclient.

Ton Voon ton.voon at altinity.com
Sat Sep 29 23:46:48 CEST 2007


Hi Alessandro,

On 29 Sep 2007, at 17:13, Alessandro Ren wrote:

>    As of the last version of nsclient, it had a check_nt_new that was
> sent with the package and it never made into the official check_nt.
>    Now that's my question, can a add new features to the check_nt code
> and submit it  to you or should we make a new check_nt?
>    We would very much like to be on the official nagios plugins  
> package
> and to contribute these new features to the community.

A similar request was made by Anthony Montibello back in March:  
http://thread.gmane.org/gmane.network.nagios.plugins.devel/4742/ 
focus=4739

And my guess is that your check_nt_new is not compatible with  
Anthony's check_nc_net.

check_nt's protocol has become a defacto standard because it is  
distributed with Nagios Plugins (both your server software support  
it) and I don't want to fragment it further without a clear direction.

If we accepted both your plugins in the core code, the ground is set  
for anyone else's special implementation of a windows agent to be  
included, and I don't particularly want this.

So I propose this: you and Anthony (and anyone else interested) get  
together and agree on a communication protocol that both your server  
software will accept. You publish the protocol on some website  
somewhere and agree to maintain that document. You may want to raise  
a ticket with http://www.iana.org/ to get an official port number  
assigned to your protocol (I managed to get one for a planned piece  
of software for Altinity called Opsview Envoy - this will be used  
internally, so we haven't published a protocol for it. You can see it  
at http://www.iana.org/assignments/port-numbers).

You can then both implement server software that takes requests from  
clients that conform to this protocol. I believe Anthony's is written  
in .NET and I guess yours is in C, but that will not matter to the  
client.

Then, I would be more than happy to accept a plugin into the core  
distribution that conforms to the protocol. I would even say that we  
can continue the maintenance of the plugin (as long as you don't  
abuse the protocol by making continual changes). For instance, if you  
both agree to support encryption later, or another authentication  
method, I can see that we would help with updating the plugin. After  
all, we maintain the check_http plugin even though we have no control  
over the HTTP protocol.

Is this fair?

Ton

http://www.altinity.com
T: +44 (0)870 787 9243
F: +44 (0)845 280 1725
Skype: tonvoon






More information about the Devel mailing list