So as I understand it if a Cluster Master goes down it brings down the entire cluster, I am currently going through the planning phase of our hosted solution setup and my major worry is that if we build our cluster and the CM goes does down it means EVERYONE goes down OR if there is a network level outage that will also bring down the enter cluster.
What do you guys do in terms of redundancy? my though was to setup multiple clusters (or its really actually load balancing) over different data centres. So at least should one DC go down or one CM go down the customers on the other cluster are not affected.
In our current clustering setup, you are right – if the cluster manager goes down, so goes the cluster. The best way at present to mitigate this risk is the same as you would any single server – backups, rsync, etc.
However, we’re currently testing a High Availability Load Balancing configuration that would still allow web traffic to be served through a backup load balancer if the CM goes down. It’ll definitely be posted here in the forums when that beta is opened up to the public
Subscribing to this thread, quite interested in this from our initial setup of Interworx. Standus, please let us know here if you hear back from them. Thanks.
Unfortunately, the InterWorx High Availability Cluster Manager beta is presently closed. We’ll be releasing a general notification of when it’s open once more. Thanks for your interest!
