At the end of February 2017 FatLab finished migrating all website databases hosted on our main configuration to a new High Availability MySQL platform with our hosting partner Rackspace.
What is High Availability
Quite simply “high availability” means that each database has 2 replicas and if one of those were to have a problem the other two remain online and the hosted websites are unaffected… If two of the databases were to have a problem, it still leaves one up and our websites running without issue…. and if all three have an issue… well the whole point is that the odds of this are tiny, very tiny.
The short of it is that our systems allows our hosted clients to rest easy knowing that critical systems such as databases are running in triplicate and any technical issues we may experience will most likely go unnoticed by them or their end users.
One of the great advantages of running a high availability system like what we have here at FatLab is that we don’t just have to leave 2 copies of every database sitting there waiting for something to happen. Instead we utilize all this power to speed up our websites in their day to day operations by load balancing read requests across all three copies. This helps ensure that busy sites remain fast and responsive as the load is shared by not one database server (such as the case in a “standard” hosting setup) but by 3 servers.
A Great Addition to Our Already Load Balanced (Redundant) Web Servers
For years now FatLab has run a load balanced web server configuration. This basically means that more than one server serves each website. Just like the databases, if one server goes down then the others are able to keep our client’s websites up and running while repairs are made to the problem server. The majority of issues we have had over the last 5 years have gone unnoticed by our clients and this is why we can measure our total down time in minutes spanning many years.
The addition of the high availability database system compliments our already redundant web servers ensuring that we provide our clients the maximum amount of uptime possible… and if our infrastructure has a problem it doesn’t also become our client’s problem.