Recommended high availability strategy?

April 15, 2013 7.1k views
So, hardware fails every once in a while. What is the DO best practice for handling this? At the moment I can only see: 1) Set up a distributed filesystem across multiple DO nodes, ie Gluster (means high latency for writes) 2) Use DNS round robin and/or failover. Problems with browser and proxy caching. But perhaps in the near future you will support: * Floating ip * Floating VM images Any other options? Thanks
3 Answers

Hey friend!

You might be happy to hear that we do have floating IPs now. You can find some information about it here:

Certainly the conversation for achieving true failover and redundancy is a larger one, of which this is merely a piece. However, having that piece is something that I think is extremely important. It lays a lot of the ground work for creative solutions on the customer side.


In the future we will be supporting something similar to a floating IP but not a floating VM and we don't have a clear ETA on that.

Otherwise HA really requires a more in depth conversation about what services you are running and what kind of availability and fault tolerance you want to build in.

I think the point I was making is the cost of purchasing any one of the vehicles you listed helix jump

Have another answer? Share your knowledge.