Mike Lindsey wrote:
> We're using an external script to handle connection caching for NSCA..
> Once we have X seconds of OCSP and OCHP results queued in the local
> cache, we fire off a batch submission through send_nsca. One TCP
> connection, one encryption handshake. Under extreme load, we've been
> having some issues and I'm thinking about re-engineering it.
>
> Only, I think if I do much work, I'd like to add the cache logic to the
> send_nsca binary itself.
>
> Changes would involve adding a cache directory and cache age (and/or max
> cache items) config directives. Without the new directives, program
> flow would be as normal. With those directives, any submitted OC*P
> events submitted to send_nsca would get dumped into the cache directory
> until the oldest file exceeds the max cache age, or the number of items
> exceeds the max. Once either of those are reached the next send_nsca
> call would submit all results at once.
>
> A lockfile mechanism of some sort would be in place to prevent any
> subsequent send_nsca runs that get kicked off, before the submitting run
> has finished, from doing duplicate submissions.
>
> Any concerns, requests, or gentle guidance towards alternate solutions?
you could look into the ndomod mechanism, when it receives a socket
timeout/blocking. therefore writing a local cache file, and resending
that once the connection is alive. this might require changes though, as
send_nsca is a single call, not a daemon to be triggered more often.
>
--
DI (FH) Michael Friedrich
Vienna University Computer Center
Universitaetsstrasse 7 A-1010 Vienna, Austria
email: [email protected]
phone: +43 1 4277 14359
mobile: +43 664 60277 14359
fax: +43 1 4277 14338
web: http://www.univie.ac.at/zid
http://www.aco.net
Lead Icinga Core Developer
http://www.icinga.org
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]