Hacker Newsnew | comments | show | ask | jobs | submit login

Surprised that nobody has mentioned replication. MySQL replication is a known quantity. PostgreSQL has it, but they're third party and less proven/battle-hardened.

That definitely is important. MySQL has fairly turn-key and easy to understand replication and has had it for a while.

The Postgres crew hides behind the "replication means different things to different people, so it would be quite presumptuous of us to build it!" mantra. It's quite annoying.

You can hack together crude replication in Postgres with Write-Ahead Log (WAL) shipping. The config has some hooks in it to automate this process and bind to it. But that doesn't allow you to do circular and/or master-master replication; the receiving node has to continuously be in recovery mode.


I wouldn't say slony is less proven or battle hardened. Being one of the very few postgres replication solutions that actually work it is currently in use pretty much everywhere where replication is required.

I think slony is less popular than mysql replication because it is a royal pain the ass to setup and maintain. In mysql you flip a switch and have replication. It has a few known issues but is "mostly" reliable and understood.

In postgres/slony you enter the wonderful world of triggers and several layers of magic.


Applications are open for YC Winter 2016

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact