Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I'd be all over pgsql if replication wasn't such a mess. Yes it works (presumably), but I can't find any useful (e.g. informative) information on a comparison between (the many??) methods to go about it and/or most modern way to accomplish it.

If anyone here knows of a good source, I'd love to see it :D (If you use the word "just" or link me the docs, you're dead to me).



Why do you reject the documentation? It has exactly what you're asking for:

https://www.postgresql.org/docs/current/static/different-rep...

It describes the various methods that are available, it lists implementations of each, it explains the tradeoffs involved, and so on. Table 25-1 summarizes the information.

Regardless of the database system being used, replication is just inherently complex. There isn't really a one-size-fits-all solution. The method to use depends on the requirements and context of a given implementation.


I think you might want to watch this tutorial from january 2015: https://www.youtube.com/watch?v=GobQw9LMEaw from Josh Berkus.

It starts with a comparison of the different ways available for replication.


Calling it a mess is unfair. It's actually quite straightforward, and the documentation (see sibling comment) guides you through the pros and cons.

For example, if you run with streaming replication — the recommended approach for most setups — then you can decide to enable WAL logs or not. If you only want replication, and not point-in-time recovery, then just running with WAL-less streaming replication is extremely simple and straightforward.




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

Search: