MySQL Performance for DevOps
summary
Shows concrete tradeoffs: every transaction with innodb_flush_log_at_trx_commit=1 forces an fsync that caps throughput to disk sync rate, changing it to 0/2 multiplies transactions per second but risks replication consistency; moving from spinning disks to flash speeds commits but exposes fsync and filesystem choices that can corrupt data if misconfigured. Demonstrates how hitting open_files_limit and too many connections stalls query parsing and open_table operations, and gives specific wins from sizing buffer_pool_size and sort buffers to cut disk reads and lower latency.
AI-written summary · about
chapters (14)
- 0:00MySQL performance intro
- 1:20Connections and queries
- 3:14Key performance factors
- 6:19Buffer pools and storage
- 8:03Memory and swap settings
- 10:15Open tables and file limits
- 13:12Spinning disk defaults
- 15:21Replication and transactions
- 17:19Concurrency and execution
- 19:45Network impact
- 21:25Query optimizer and stats
- 23:20Indexing strategies
- 26:03Sort and join buffer tuning
- 27:32Performance tuning tips
