You use a clustered IP solution like heartbeat. This allows for a secondary machine to assume the clustered IP address and become the primary HAProxy load balancer.
From: Benjamin Scherrey [mailto:scherrey#proteus-tech.com]
Sent: Tuesday, March 11, 2008 10:08 AM
Subject: Resiliency of HAProxy in the Datacenter
Pardon the possibly naive question but I'm a services programmer & not a systems infrastructure guy. What is the best practice for preventing my haproxy machine from becoming a single point of failure? Do I deploy it in pairs and, if so, what's the fail-over detection/implementation mechanism? Appreciate someone giving me a clue as to how this is done.
This archive was generated by hypermail 2.2.0 : 2008/03/11 17:30 CET