Hi there,
We're looking into setting up a production ES cluster on EC2 (thanks to the
various guides/plugins, it's pretty easy to setup). EC2 has had it's fair
share of publicized outages so for HA reasons, we would like to maintain
our indices outside of ES also. We would be maintaining these indices
across a number of regions if time/data sync performance allows
The options I see are:
a. Use gateway/other backup techniques to write to an S3/external data
store and piggyback from that. (almost like a separate cluster)
b. Maintain a separate non EC2 slave node (which would become the master
on catastrophic EC2 fail?)
Does anyone have any experience of doing something like this? Would really
appreciate any tips/advice on this and also on experience with cluster
setups on EC2 across multiple zones/regions.
Thanks in advance,
Derry
--