SmartThings First 8 Years with Cassandra
summary
Eight to nine years of production Cassandra: more than 300+ nodes across over ten clusters with two global rings. Operations migrate from DataStax Enterprise to Apache Cassandra, replace fragile homegrown backups with Medusa backup, adopt Reaper for repairs, move from direct-attached storage to EBS-backed volumes, and split noisy multi-tenant rings into single-app clusters. Application changes include driver protocol pinning to avoid rollout bugs, avoiding SELECT * after schema edits, and shipping a time-window compaction JAR across nodes.
AI-written summary · about
chapters (20)
- 0:00SmartThings and Cassandra intro
- 1:26Early Cassandra architecture
- 3:33Backup and monitoring early
- 5:16Developer opinionated libraries
- 7:03Monolith to microservices
- 8:05Scale issues Super Bowl
- 9:10Switching to Cassandra
- 11:17Expanding into China
- 12:20Migrating to EBS volumes
- 14:04App monitoring driver best
- 16:35Redatacenter era optimizations
- 18:38Data center loss Reaper
- 20:03Offline-first multi-region
- 21:48Pandemic operational shifts
- 22:12Present day microservices
- 23:16Modern backup Medusa
- 24:21Deprecating monolith lessons
- 25:25Eight years Cassandra lessons
- 28:16Cassandra key takeaways
- 29:50Q&A closing
related talks
46mApache Cassandra: 8 years in production
35mReplacing Cassandra: A Digital Transformation for the World’s Largest Digital Identity Network
29mCassandra - an introduction
53mApache Cassandra. Inception - all you need to know
23mDeciphering Cassandra’s Prophecy: Preventing the Fall of Your Application
35mMission Critical Data, and Evaluating the Community in OSS Tech
24mZero Downtime Mastering Cassandra Live Migrations at Scale
46mCassandra Serving Netflix @ Scale