Database Replication and ClusteringΒΆ

Aptible Deploy supports replication or clustering for a number of Databases:

  • Redis

    Aptible Deploy supports support creating read-only replicas for Redis.

  • PostgreSQL

    Aptible Deploy supports read-only hot standby replicas for PostgreSQL databases. PostgreSQL replicas utilize a replication slot on the primary database which may increase WAL file retention on the primary. We recommend using a Metric Drain to monitor disk usage on the primary database.

  • MySQL

    Aptible Deploy supports creating replicas for MySQL databases. While these replicas do not prevent writes from occurring, Aptible Deploy does not support writing to MySQL replicas. Any data written directly to a MySQL replica (and not the primary) may be lost.

  • MongoDB

    Aptible Deploy supports creating MongoDB replica sets. To ensure that your replica is fault tolerant, you should follow the MongoDB recommendations for number of instances in a replica set when creating a replica set. We also recommend that you review the readConcern, writeConcern and connection url documentation to ensure that you are taking advantage of useful features offered by running a MongoDB replica set.

Replicas can be created for supported databases using the aptible db:replicate command.

Warning

Creating a replica on Aptible Deploy has a 6 hour timeout. While most databases can be replicated in under 6 hours, some very large databases may take longer than 6 hours to create a replica. If your attempt to create a replica fails after hitting the 6 hour timeout, reach out to Aptible Support.

When deployed in a replicated or clustered fashion, Aptible Deploy will schedule Database Containers across different AWS Availability Zones to ensure high-availability.