Thanks for your answer.
How do you manage the failover with your virtual servers ?
Postgres has a built-in failover mecanism using the trigger_file setting in reconvery.conf.
Do you handle it manually or use another system to automate it ?
Robin.
De : Poul Kristensen <bcc5226@gmail.com>
Envoyé : jeudi 6 avril 2017 18:16
À : Robin LUCBERNET
Cc : pgsql-admin@postgresql.org
Objet : Re: [ADMIN] How do you manage cluster replication and failover ?
The challenge is not whether to use the difference tools as you mentioned. The challenge is to make to optimal database design the get the optimal backup/restore or disaster/recovery.
If you have just one database per Postgres installation and then just handle user/role security using schema's in one database then it is a good choice in my opinion as only one installation to maintain.
If you have multiple databases I think that you have to be aware of issues concerning the recovering of one database in an easy way. Furthermore be aware of the possibilty of corrupted databases. A regularly dump will help. If using Debian's hdparm -I /dev/.. is usefull on Debian.
Write performance: I have succeded writing 10 mill. records in 50seconds using 2 virtual servers with streaming replication. I haven't yet tried physical servers yet.
Hope this is usefull.
BR
Poul