Hacker News new | past | comments | ask | show | jobs | submit login

Lots and lots of replication, more or less. I don't know how it works in Postgres, but with something like Mongo, you set up a replication cluster and presume that up to (half - 1) of the nodes can fail and still maintain uptime. Postgres, being a relational database rather than a document store, likely has an additional set of challenges to overcome there, but it's very possible to do.



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: