Multi-master replication for painless fault tolerance by Stas Kelvich

Tuesday, November 15 at 11:20-12:15

We'll cover following topics:

* State of cluster and replication solutions for postgres and niche for multi-master in that ecosystem
* Some theory about distributed databases design and tradeoffs
* Raft storage as postgres extension for node discovery
* Fault detection and recovery in multi-master
* Framework for testing multi-master behaviour with respect to network falures and splits

Also will show live demo of node recovery with multi-master on small Raspberry Pi cluster.

About the Speaker

Started hacking multidimensional indices in postgres back in 2012. Since then worked also on full text search, two-phase commit, logical decoding, distributed transaction.

Also wrote a lot of ruby on rails code and numerical computation code for laser physics simulations.

 

Tuesday, September 13, 2016 - 10:00