Ilya Kosmodemyansky at RIT++ 2012

on YouTube ↗10m

" summary

Discusses real production problems encountered with MongoDB: a master-election bug caused cluster collapses and unpredictable downtime, and journaling was added late. Argues lack of true transactions leads to integrity bugs in money/accounting systems, naive two-phase locking is incorrect, and optimistic concurrency forces roughly one server per writing thread; recommends migrating back to SQL to reduce server count and improve reliability.

AI-written summary · about

" related talks

• 34mPanel Discussion: DatabasesDenis Reznik, Andrey Varenyk, Alexander Solovyov, Nedashkivskyi Yevhen · Highload fwdays'19
• 41mTransactions: myths, surprises and opportunitiesMartin Kleppmann · Strange Loop 2015
• 29mModern Database Architecture: From SQL to NoSQL & Distributed SQLAndrei Manakov · Site Reliability Engineering (SRE) 2025
• 03mKristina Chodorow interviewed at OSCON 2010Kristina Chodorow · OSCON 2010
• 01h 08mMongoDB in Production — Myth or Reality?Alexey Tokar · JavaDay Kiev 2014
• 10mEverything You Know About MongoDB is WrongMark Smith · PyOhio 2020
• 47mBeing transactional in 2019 - what does this actually mean?Maciej Próchniak · JDD 2019
• 43mBeing Transactional in 2019 — What Does That Actually Mean?Maciej Próchniak · Confitura 2019