"Beauty and the Beast: Applications and Distributed Systems"
summary
Defines distributed systems with concrete examples: Elasticsearch index creation with wait_for_active_shards=2 returns shards_acknowledged true then false when a node dies; MongoDB uses sessions with causal_consistency=true and atomic update patterns (addToSet + match-and-increment) to make retries safe; Elasticsearch uses document version optimistic locking for idempotency. Recommends retry once for transient network blips, backoff for outages, avoid retrying authorization errors, tune client options like heartbeat_frequency=10s, and test on real clusters by killing nodes.
AI-written summary · about
chapters (13)
- 0:00Introduction and speaker background
- 3:12What is a distributed system?
- 5:19Consensus in distributed systems
- 7:26Synchronization and logical clocks
- 8:50Election algorithms and roles
- 10:15Replication and consistency models
- 14:10Fault tolerance and redundancy
- 15:13Errors and retrying in applications
- 18:29Categorizing errors for retry logic
- 22:03Idempotent operations and safe retries
- 25:14Understanding client gem APIs
- 27:19Testing against distributed systems
- 28:43Real-world application and Q&A
related talks
48mBeauty and the Beast: your application and distributed systems
44mDistributed systems and scalability rules
25mGrowing Distributed Systems
45mDo Distributed Systems they said. It'll be fun they said.
02h 18mDistributed Systems: In the Daily Life of the Data Engineer
01h 00mDistributed transactions implementation trade-offs
52mDistributed Computing Dealing with Time and Failure in the Wild
41mIntroduction to distributed systems