Does scidb support standby instances for the coordinator?


If it does, is it a hot one or a cold one?



Not yet.

There are a couple of “single points of failure” in the current implementation:

  1. At the moment, you can specify only one SciDB instance to be the “coordinator”, and direct all client connections to it. We’re going to be making it possible to run multiple coordinates in an upcoming release.

  2. The PostgreSQL DBMS is our catalog and plays an important role in the transaction management apparatus. In a future release, we’re going to be pulling both the system catalog meta-data, and the transaction management, entirely within SciDB, and use SciDB machinery to organize it.

You’ll have to bear with us for a while. Neither of these projects is difficult. They’re just lower priority–at the moment–than improving what we have in the core feature list: the engine’s performance, scalability, and robustness.