MySQL (NDB) Cluster Best Practices
summary
MySQL Cluster (NDB) is an in‑memory, sharded ACID store with synchronous replication via a linear two‑phase commit and checkpointing to disk; a betting customer pushes 8 million ops/sec on four nodes, HDFS metadata moved into a 12‑node NDB cluster scales name‑node throughput. Config is via config.ini, data nodes mirror in two‑node partitions (if both fail the cluster stops), and practical limits are up to 48 data nodes. Tune threads and buffers (example knobs: redo buffer=64 long message buffer=128 scans=500), use SSDs and many cores, and perform rolling restarts for config changes.
AI-written summary · about
chapters (14)
- 0:00MySQL NDB Cluster intro
- 1:48NDB Cluster history
- 6:04Use cases and real-world applications
- 15:08NDB Cluster architecture
- 20:09Data partitioning & replication
- 27:55Optimizing data nodes
- 33:37Configuration & hardware
- 39:51Rolling restarts & ops
- 43:50Monitoring tools
- 44:53App design best practices
- 47:24NDB vs InnoDB
- 49:54NDB vs Galera
- 56:37Network partitions & scaling
- 58:09Benchmarking & performance
related talks
33mMySQL Cluster as Transactional NoSQL
37mMySQL Database Scalability
24mFlinkNDB : Skyrocketing Stateful Capabilities of Apache Flink
49mOnline Database Reliability, Performance, and Consistency Engineering
25mThe future of databases: Combining operational and analytical data at massive scale
27mDistributed MySQL Architectures: Past, Present and Future
20mExtreme Performance and SmartScaling MySQL with Storage Class Memory Servers
53mMySQL DevOps