[Nagiosplug-help] Monitoring a LVS direct routing cluster

[Thorsten Tüllmann]

Hi there,

is there an easy way to monitor a Linux Virtual Server cluster with Nagios?

We're using direct routing 
(http://www.linuxvirtualserver.org/VS-DRouting.html) to load-balance our 
web servers. To monitor the virtual hosts on the real servers it is 
necessary to connect to them directly on layer 2 with the virtual IP all 
servers share.

The straight-forward way to monitor such a setup would be writing a 
check_http which can address hosts by their IP _and_ MAC. The problem is 
such a plugin does not exist (or I could at least not find one). I 
couldn't even find a socket program that allows TCP connections to a 
specific MAC and IP. The only way seems to be using raw packets and I 
don't want to reimplement a TCP client with libnet or something similar.

My next best solution is manipulating the ARP cache of the monitoring 
machine to connect to the real server one wants to check. This is quite 
an ugly hack and can possibly cause serious trouble.

So here I'm running out of good ideas. Is someone out there monitoring 
LVS clusters with Nagios?

TIA & Regards,
Thorsten