PostgreSQL Replication with Easy Failback

Рет қаралды 12,219

Күн бұрын

To get the commands used in this tutorial, visit: www.scalingpostgres.com/tutor....
Learn how to failover to a replica database and then failback to the original primary database easily without a full restore of the database and without using pg_rewind.
This tutorial assumes you already have streaming replication setup using replication slots, if not, check out these two videos:
• PostgreSQL Streaming R...
• PostgreSQL Replication...
Want to learn more about Postgres performance?
Join my FREE mini-course called the PostgreSQL Performance Starter Kit here: www.scalingpostgres.com/cours...

Пікірлер: 9

@decipheringthematrix56 5 жыл бұрын

Nice! I can't wait to try this.

@CR-rg5nh 5 жыл бұрын

Hi! Do i need Replication slots for this situation (failback) or i can do it without slots? Thanks

@adammulla9181 4 жыл бұрын

Hi, Thanks for your tutorial.. Can we do switchover and switchback without adding parameter recovery _target_timeline='latest' in 9.4 My friend told me he has done switchover and switchback in 9.5 without adding recovery _target_timeline=latest. is it possible ?

@1487atul Жыл бұрын

I think there is no need to define restore_command in recovery.conf file if you are already creating a replication slot. please correct me if I am wrong as far as I know restore_command is needed when you are archiving xlog file to archive location but if you have replication slot then there is no need of restore_command . please correct me If I am wrong ?

@saintsun90 3 жыл бұрын

Hello. Its not working v12 do you have any solution about it

@mrmik82 3 жыл бұрын

Version 12 doesn't use recovery.conf but recovery.signal, standby.signal and postgresql.auto.conf.

@davebrunkow9194 5 жыл бұрын

Everything works great until I remove the replication slot. I immediately get replication terminated by primary server. Only a pg_basebackup gets me out of it.

@davebrunkow9194 5 жыл бұрын

I did not add recovery_target_timeline = 'latest' to the recovery.conf on the replica which caused all kinds of funkiness; WAL streaming at LOG: replication terminated by primary server DETAIL: End of WAL reached on timeline. DOH!