We use amazon S3 buckets for publishing js files and putting them on client web sites. In other words, js files are stored in amazon s3. The main problem is since these files are on client sites, we don’t want to give the client new js codes if amazon s3 bucket goes down. Is there way in a Rails app to do the redundancy switch without giving client another code to install on their website? We want that piece of code to go use another bucket if the other one goes down without republishing anything or giving a new js code to the client. It seems impossible but I thought somebody might have an idea. Thanks in advance for all your help.
Share
I’ve never heard of an extended S3 failure affecting a single bucket. Indeed, I’ve only heard of one large-scale read outage in its history. Routing issues and the like might give you downtime, but changing buckets won’t fix that, since they’re all serviced by the same frontend.
If you want to have a backup plan, one good option is to use a CNAME as Rahman suggests, but your backup system should not be on S3; load it from an ordinary VPS on some other provider (not EC2). Note that the old CNAME value will be cached for its TTL; setting a very long TTL will mean it will take longer to failover.