Lxd failover high availability

LXD
#1

Hi All,

I hope that someone can help me. I now have a system with btrfs raid 10 to store my containers and vm machines.

I have a spare system that i want to use as a failover for the first system. So when the first system goes down i want the second to automatic take over.

I dont have shared storage. so i want to go with local storage. is this possible? if so what doe i need to make the above happen?

Also how do we configure the ip’s? because it has to be the same ip adresses i think?

What is the easy way to accomplish this?

Kind regards,

Chris

#2

If you’re not using shared storage then you’d need to setup failover/high availability for the services inside your instances using duplicated instances on each cluster member.

Then for IP failover, this depends on the networking mode you’re using, but for bridged networks you can have your instances advertise a floating IP using something like keepalived.

So effectively LXD doesn’t really get involved with making your services highly available, it just allows you to continue managing the instances that are not on the member that is down.