Hello,
Since most of the threads in this forum discussing failover and disaster recovery (DR) solutions are several years old, I kindly ask not to be referred to a Google search (I already did it).
I am looking to implement an active-active disaster recovery solution for DirectAdmin. The goal is to ensure that all data is replicated across nodes, and in the event of a failure on Node A, all traffic will seamlessly be routed to Node B.
Here is the setup I am considering:
Thank you!
Since most of the threads in this forum discussing failover and disaster recovery (DR) solutions are several years old, I kindly ask not to be referred to a Google search (I already did it).
I am looking to implement an active-active disaster recovery solution for DirectAdmin. The goal is to ensure that all data is replicated across nodes, and in the event of a failure on Node A, all traffic will seamlessly be routed to Node B.
Here is the setup I am considering:
- 2 DirectAdmin servers
- 3 SQL servers (Galera Cluster/Group Replication)
- 2 ProxySQL instances to manage read/write traffic distribution across the SQL cluster
- 1 SQL load balancer using round-robin to distribute traffic between both ProxySQL instances. DirectAdmin will be connected to this load balancer.
- S3-compatible solution (e.g., GCP, AWS, etc.) to store all user data, ensuring accessibility from both nodes.
- Does DirectAdmin support external DNS servers, such as PowerDNS, for syncing DNS records and facilitating easier switching in case of a failure?
- How can I synchronize users, emails, and FTP accounts effectively? Using rsync for /etc/passwd, /etc/group, and similar directories feels more like a workaround rather than a reliable solution.
- Is there anything else I should consider in this setup?
Thank you!