Daniel Mewes

Thanks for the information @digipigeon . We have seen instances of data corruption on OpenVZ before (see https://github.com/rethinkdb/rethinkdb/issues/4750 ), so this is probably related. Did anything unusual happen before this...

Until we have figured out what exactly causes the corruption, we should try to detect OpenVZ environments and print a warning on startup.

I know it has been a while, but were you using secondary index queries? We have just fixed a bug in secondary index construction that could lead to corrupted secondary...

Thanks for the pull request @marshall007. I think this is your first C++ PR? :-) I'll take a look at it later.

I added a few comments, but looks good mostly. Are you planning to add support in the other drivers as well? We can take over that part if you prefer.

Just saw that there was some discussion in https://github.com/rethinkdb/rethinkdb/issues/3148 on whether this option should be called `stride` or `step`. I personally like `step` better. I'll ask a few people around...

Thanks for updating the PR and for the `isPlainObject` refactor. Let's keep disallowing `step == 0`. It avoids potentially unexpected behavior and allowing something later is always easier than the...

@marshall007 After talking to some people, we settled on `step`. The server side of this looks good to me. I haven't reviewed the JavaScript driver changes yet.

I'm not sure why this is happening, but does it work when you retry after installing the eventmachine gem manually?

@dminkovsky Thank you for chiming in here. I'm interested in your opinion: Would you prefer the DNS based solution over a direct integration with etcd? Does anyone know how other...