[Nagios-devel] NSCA rewrite ?
Posted: Wed Feb 01, 2006 2:20 am
On Tuesday 31 January 2006 19:00, Ethan Galstad wrote:
> I'm planning a rewrite of the NRPE and NSCA addons to support multi-
> line plugin output that should appear in Nagios 3.x. In doing this,
> I was thinking that communication between the NRPE/NSCA servers and
> their clients should be done using SOAP or XML-RPC. Any thoughts on
> this?
There is one other issue with NSCA. NSCA uses TCP connections at the moment. I
believe most of the NSCA environments are using NSCA as obsessive compulsive
processor command. If the machine (or the nscad) to which commands are
transmitted crashes, nagios gets blocked (serial execution) for every command
that is being checked until the command processor timeout is reached.
Imagine an environment with more than 1000 services and a ocsp timeout of 5
seconds. All checks are now executed about every 83 minutes.
We solved this problem with a wrapper around the NSCA client, which checks for
a local file created by keepalived, to ensure the backup system (nscad) is at
least alive.
I would prefer some client / caching daemon on the client side. If you're
about to rewrite the NSCA utils, maybe you can solve this problem too.
I hope i'm not alone with this ..
Best regards
Percy Jahn
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]
> I'm planning a rewrite of the NRPE and NSCA addons to support multi-
> line plugin output that should appear in Nagios 3.x. In doing this,
> I was thinking that communication between the NRPE/NSCA servers and
> their clients should be done using SOAP or XML-RPC. Any thoughts on
> this?
There is one other issue with NSCA. NSCA uses TCP connections at the moment. I
believe most of the NSCA environments are using NSCA as obsessive compulsive
processor command. If the machine (or the nscad) to which commands are
transmitted crashes, nagios gets blocked (serial execution) for every command
that is being checked until the command processor timeout is reached.
Imagine an environment with more than 1000 services and a ocsp timeout of 5
seconds. All checks are now executed about every 83 minutes.
We solved this problem with a wrapper around the NSCA client, which checks for
a local file created by keepalived, to ensure the backup system (nscad) is at
least alive.
I would prefer some client / caching daemon on the client side. If you're
about to rewrite the NSCA utils, maybe you can solve this problem too.
I hope i'm not alone with this ..
Best regards
Percy Jahn
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]