Auto database mirroring.

[M Hadi Azarabad]

Hi. I'm looking for a solution to save data on mirror database. Using database routers allow us to save a record and lets us decide which database must be used and I think manual database selection is not optimal because it uses ORM to create raw SQL twice or more (based on the number of mirror database).

model_obj.save(using='db_1') 
model_obj.save(using='db_2') 

I think django can handle this problem and optimize it by adding setting list variable called MIRROR_DATABASES so all mirror databases place in there and Django save all of records by creating one raw SQL if mirror databases using the same backend.

[Mariusz Felisiak]

As far as I'm aware database mirroring is not something a web framework should be responsible for, see a ​handy comparison of different solutions in PostgreSQL's docs. Duplicating all the operations would be inefficient, very complex, and error-prone.

[Steven Mapes]

Following on from Mariusz's reply with the Postgres reference on HA, load balancing and replication here's ​MySQL's replication solutions and ​MySQL Proxy which can be used as load balancer to distribute requests and perform automatic failover, here's a ​Digital Ocean install guide for the latter.

If you are using MariaDB then here's the ​replication docs to get you started.

If you are looking at Multi-Master using either then look into ​Galera Cluster but these will be better solutions than writing twice from code