16.9 and Instance Stability


SciDB 16.9 is available. Please review the release notes carefully. It’s important to highlight a particular change to instance lifetime and stability.

In 16.9 the query synchronization architecture was changed significantly and a new clocking mechanism was added to EE to allocate threads for query execution in a more efficient manner. As a result, the Community Edition (CE) does not recover as cleanly from the crash of an instance. In the rare event that an instance crashes, the full SciDB cluster may need to be restarted. This problem only affects the CE starting with 16.9. The Enterprise Edition provides machine level replication and automated instance failure recovery.


this is the reason cerbo.io forked scidb source code and adding patches to the fork to prevent issues like that to production systems which are running community edition code.


we also accept patches from community here https://github.com/cerbo/scidb