Running a Cloud-Native SQL Database on Mesos
summary
CockroachDB runs on Mesos/DC/OS using the DC/OS SDK and installs with a single command to launch a three-node cluster that can be resized live (demo expands to five nodes). It uses Raft-based three-way consensus replication with serializable transactions, automatic rebalancing of 64MB ranges, and survives a data‑center loss with two‑of‑three quorums. Enterprise adds fast incremental backup/restore; single‑node performance is roughly within 2x of Postgres.
AI-written summary · about
chapters (16)
- 0:00Start
- 3:23Weakly Persistent Local Storage
- 6:52Distributed Execution of Your Sequel Queries
- 7:28Flexible Scheduling
- 8:29Multi Active Availability
- 9:50Active Active
- 13:09Distributed Transactions
- 14:10Demo
- 17:05The Replicas per Node
- 21:07Current Status of Cockroach
- 21:43Distributed and Incremental Backup and Restore
- 23:17Csv Importer
- 25:14Where You Can Find Out More about Us
- 29:00Single Node Performance
- 29:29Impact of Synchronous Replication
- 36:37Row Level Partitioning
related talks
48mBuilding mission-critical applications with CockroachDB
24mNavigating Distributed Databases: Challenges & Opportunities
01h 11mFearless global transactions with
52mCockroachDB
42mThe hows and whys of a distributed SQL database
31mFuture-proofing operational data at scale
14mHow CockroachDB can enable Multi-Region Deployments and Data Locality 
37mBuilding a Cloud-Native SQL Database