Make your failover latency predictable | Architectural Simplification

Description

When it comes to failovers, distributed SQL database CockroachDB requires no human intervention. Nodes simply repair themselves, rebalance data, all to survive outages and create predictable behavior for users of that database. Join as Cockroach Labs Technical Evangelist Rob Reid: ✔️ Simulates a severe production failure whereby a whole cloud region has become unavailable. ✔️ Compares CockroachDB vs Postgres when it comes to survivability. ✔️ Shares why having a cluster of CockroachDB nodes may be preferable to having a primary and secondary failover setup. ✔️ What you should know about RPO and RTO. Make your failover latency predictable | Architectural Simplification 00:00 Introduction 00:14 Diagrams of failover setups 00:40 How does replication work in CockroachDB? 01:35 Creating a Postgres primary and secondary setup 03:14 Taking down a primary node in Postgres 03:42 The primary failover scenario was successful... but now what? 05:12 How does CockroachDB handle a failover scenario? 07:50 A few notes on our testing approach 08:15 What did this testing prove? Additional resources: ✔️ What is CockroachDB? https://cockroa.ch/42uaZ9G ✔️ CockroachDB architecture overview: https://cockroa.ch/3RzIvrh ******************************** Follow Cockroach Labs ✔️ Website: https://cockroachlabs.com/ ✔️ Twitter (X): https://twitter.com/cockroachdb ✔️ Linkedin: https://www.linkedin.com/company/cockroach-labs/ ✔️ Facebook: https://www.facebook.com/cockroachlabs ✔️ Instagram: https://www.instagram.com/cockroachdb ✔️ Threads: https://www.threads.net/@cockroachdb