I can not tell you how pleased I was to discover HAProxy. What a brilliant piece of kit!
At work I have 3 IIS servers that front a db backend. Historically the company used the MS NLB which has worked to a degree but does nothing but round robin requests with no ability to report node issues, let alone shoot them in the head!
So I was looking for a solution and came across HAProxy - now understand I am not a Linux guru, my exposure has in the past been very very limited so I have had a bit of a learning curve to do.
I have to say that ‘out of the box’ getting a single instance of HAProxy to load balance across the 3 IIS servers was surprisingly easy.
But you know how it is… I decided the better solution would be to have two instances of HAProxy in an ‘Active/Passive’ configuration to provide a good level of resilience.
I have followed a tutorial from ‘Clusters from Scratch’ which was really useful and another from Digital Ocean as well as dipping into many other resources.
As it stands at the moment I have Pacemaker / Corosync / PCS and HAProxy installed on 2 Ubuntu Servers. The cluster is up and running though I am yet to configure fencing and STONITH.
I am now stuck, probably because I’m not seeing the obvious. I have a VIP setup that the cluster uses as a resource, I now need to add HAProxy as a resource in Pacemaker, before I can I need to setup the HAProxy.cfg file.
In some of the tutorials it says that I need to bind the front end to the host IP, other resources say I need to bind the front end to the virtual IP.
Can someone please clarify this for me?