Strange errors under high concurrency with Sidekiq

[apa512]

NoBrainer often crashes in weird ways when I'm running too many Sidekiq jobs at the same time.

Communication with the Rethink server misbehaves in one of two ways:
can't modify frozen IOError: https://gist.github.com/apa512/61321dbb81ff529fc5f0
Server dropped connection with message: "SE": https://gist.github.com/apa512/15b8fb53b086e6a0935a
Exact message varies, but basically communication with the server seems to be cut off ("SE" should be "SUCCESS").

I can't find anyone else with the same problem and I'm not even sure if to look at NoBrainer, the Rethink client or something else. Any idea what the problem could be @nviennot ?

[apa512]

Using server version rethinkdb 1.13.3-0ubuntu1~trusty (GCC 4.8.2) and Ruby client version 1.13.0.1.

[nviennot]

I've added a connection lock, until we get a real connection pool.
So that should fix the issue.

[apa512]

Much appreciated.