Other options to consider:
- Region 1: AWS Aurora Postgres with 1 writer and 1 reader node.
- Region 2: Same as above.
- Choose a 3rd party tool or use pg-logical to replicate data between Region 1 and Region 2 so that they are always in sync.
Please note, your app tier should align with these.
Thank you
Kam
From: Mauricio Fernandez <mmauricio.fernandez@xxxxxxxxx>
Sent: Thursday, January 23, 2025 1:01 PM
To: Gabriel Guillem Barceló Soteras <gbarcelo@xxxxxxxxxxxxxx>
Cc: Pgsql-admin <pgsql-admin@xxxxxxxxxxxxxxxxxxxx>
Subject: [EXT] Re: High Availability
|
External Email: Use caution with links and attachments.
|
Thank you ver much for answering Gabriel
I think there is no gold-standard HA solution. Some commercial PostgreSQL offer HA as an add-on or standalone
product, and they tackle different SPOFs with varying components. If you got the open source way, a good starting point is
https://www.postgresql.org/docs/current/high-availability.html
Depending on your requirements you may go one route or another. For exemple, it may be ok for you a single
pgpool-2 instance, without front Haproxy tier.
We tested some of them, and we don’t use any of them so far. In our scenario we can assume some out of office
downtime for maintenance tasks.
Good luck,
Gabriel
Hi admins, I need to know if anybody know about some survey, poll, web site, etc talking about most popular postgresql's most popular high availability solutions.
Using repmgrd, with or without pgpool-ii, haproxy, etc.., commercial or free.
|
