how does work failover for static ips?

[benoitc]

I didn't test it yet but how does HA failover work if i want to keep the service live ? Does it reassign the IP address to the new running instance? Or should i still use something like keepalived? If not what does t used to do it?

 

[ph0x]

Are you talking about the hypervisor level HA of VM/CT?
The machines of a failed node get restarted on another node after a couple of minutes. They keep their normal address. Or do you have something else in mind?

 

[benoitc]

so machines are « simply » restarted? no copy of the memory?

 

[ph0x]

How would you copy the memory over from a failed node?

 

[benoitc]

well it can be copied over the time. not the last stae of course. but right...

 

[ph0x]

If you need that level of failover redundancy you should achieve it on application level, i.e. run a clustered service. The hypervisor HA failover only guarantees that the machine is running.