How not to Fail at Distributed Systems with Examples from Riak
summary
Shows concrete Riak failures and fixes: an app that SELECT suddenly pulls a user_tags comma‑list ballooned to thousands and turns one fast query into thousands of microcalls; expensive operations like key‑listing cripples production until Riak KV 2.9 adds safe listing; OS-level memory harm from disable transparent huge pages* advice (turning THP off recovers memory); mentions OTP 22 message fragmentation as mitigation for head‑of‑line blocking. Lessons focus on full‑stack troubleshooting and concrete config fixes.
AI-written summary · about
chapters (5)
- 0:0003:05 - Introduction
- 3:0611:05 - Reviewing your work
- 11:0611:31 - Just accepting the facts
- 11:3211:49 - The importance of being Earnest
- 11:5032:12 - Wrap up with QnA
related talks
34mDistributed Systems - What Can Go Wrong and How to Get It Right
29mHorrors of Distributed Systems
25mGrowing Distributed Systems
33mWhy Are Distributed Systems So Hard?
36mDesigning for failure - architecting fault-tolerant system
45mDo Distributed Systems they said. It'll be fun they said.
28mWhy Are Distributed Systems So Hard?
09mJustin Sheehy interviewed at Velocity 2011