> It is possible to mount an EBS volume and have RethinkDB store your data on it, but that setup is not available out of the box.

That's a curious default choice. It's hard for me to imagine why you'd ever want to run RethinkDB on local instance storange, instead of EBS, given that EBS is much more reliable, is easier to back up, can be swapped between instances, etc. (Or is this AMI more for just trying out, not for actual production use?)

Obviously, EBS has its problems, but so does local instance storage.

Also: I remember that way-back-when, RethinkDB was touted as being designed for SSD's from the ground up. Is that still the case? Would running on provisioned-IOPS EBS provide a huge benefit over regular EBS or regular local storage?

The documentation is misleading. RethinkDB instances launched from the Marketplace all run on an EBS volume.

For anyone reading this, the docs have been fixed. Sorry -- it was my screwup.

I know the answer to your last question which is that the SSD optimized code is all present as the serializing layer so running on SSDs will provide a significant performance boost. I'm not 100% sure what provisioned-IOPS EBS is but I'm guessing from context that it's solid state based?

No, it's just plain ol' EBS which delivers a minimum of X IOPS, where X is between 100 and 2000 (unless they have improved this since I last checked).

AWS does have some SSD-based storage, but only for the larger instance types.

x <= 4000

Thanks Jeff. I've learned to always provide the caveat "unless AWS have upgraded this recently" due to the speed at which new features are added/improved. :)

It looks like for now you need to manually edit a config file to get this working with additional EBS volumes, but the wording implies better support is coming.

I wish I could use everything. I want to use it all. All the new platforms and technologies and even old ones that are still awesome. RethinkDB, mongodb, redis, Firebase, Parse, meteor.js, Postgres, Go, etc etc. Just take all that awesomeness and put it in an app and have it do awesome stuff.

I'd really hate to taste your cooking /joke :)

I'd really hate to debug the app.

Fry it and call it okonomiyaki.

Just not omakase.

Not so much a joke as an excellent analogy.

App all the things!

Awesome - good job making it even easier to start!

That's great, but what is the strategy for data persistence?

As I understand it, if you aren't using EBS, unless you snapshot your instances, they are reset to their last snapshot once they are shut down. Doesn't that seem like a huge risk to put your database under?

Are there good strategies for mirroring data onto EBS or running a slave instance or something?

Sorry, the storage is EBS, the line in the docs is just misleading. We'll push a fix to the docs in a few mins.

Just pushed the fix live. Sorry for the mistake.

What have you done with Marc, and why is he never on Steam these days?

Marc is a live and well! I'll tell him you're looking for him :)

This is such a great feature. I'm really looking forward to using RethinkDB as it matures.

Cool, an easy way to test the clustering features.

Heroku next?

There isn't an official add-on yet, but as Kudos pointed out, RethinkDB Cloud has an add-on in beta: https://addons.heroku.com/rethinkdb-cloud

This is not something that could evolve into a Heroko add on. What you're thinking of is something like https://www.rethinkdbcloud.com/