Morning all

I have been tasked with creating a solution that load balances requests to a farm of IIS servers that is a front end of an i Series IBM db.

Traditionally, they have used in MS NLB which just uses DNS roundrobin to farm the requests out to the various components of the web farm. As a stop gap it works to a degree but has zero functionality as far as detect failed nodes etc.

So in looking for a solution I stumbled across HAProxy -I set one up and it works really really well. I'm not traditionally a Linux person so was very pleased to discover how relatively easy it was to set up.

Of course that led me down the path of then looking at a HAProxy cluster, just a bog standard Active/Passive setup, which then led me onto the excellent 'clusters from scratch' tutorial. I've really enjoyed this experience, but my experience in Linux is very limited.

As it stands using Pacemaker and PCS I have a working active/passive cluster, I have failed them over and it seems rock solid but I'm having trouble in understanding where in the process , or how even I configure HAProxy to become a 'resource'?

I have configured the haproxy.cfg with the virtual IP of the PSC cluster, which is working, but I'm guessing that HAProxy needs to become a 'resource' in order for Pacemaker to monitor the service? Firstly am I right in that statement?

If I am how do I configure HAProxy / Pacemaker to work together? Sorry for being so base in my question!

The lab setup is on Hyper-V, the Server OS is for HAProxy and Pacemaker is Ubuntu.

Any help greatly appreciated, many thanks.

Have you considered running haproxy on FreeBSD with CARP? It will probably be easier to setup than pacemaker. Just set your bind addresses in the haproxy config to 0.0.0.0 or ::/0