Is there a doc for active/passive configuration?

Looking here there are docs for active/active. But for active/passive it just says it will use DRBD. I don’t see anything in /etc/gitlab/gitlab.rb that would obviously relate to DRBD, so I guess that has to be installed and configured outside of OmniBus. I do see stuff on setting up Redis Sentinel for master/slave, within the active/active documentation. Is that just for automatic failover in an active/active configuration without a separate load balancer?

I have two VMs and big NFS volume mounted to both. I can follow the active/active doc on NFS configuration but I can’t tell if that’s adequate for active/passive (ie if it depends on the shared pgsql and shared redis instances).

Any pointers/writeups would be appreciated.

Does active/passive require setting Redis in master/slave configuration? What about SQL master/slave replication? Or does Redis master/slave handle keeping the separate databases correct?

I didn’t receive any tips on Gitter or IRC either, so I opened a issue to provide better documentation

The basic questions were answered in the ticket but I would still like to see the published documentation improved. If you agree please thumbs-up the issue.

To summarize: The entire /var/opt/gitlab should be shared between instances, only one instance should be running the application at any given time, and failover process it simply to shutdown GitLab on active server (if it can be reached), and startup GitLab on passive server (making it the active server). Some enforcement mechanism to ensure that service is not started on both machines at same time, or even further ensuring shared file system is only writable from active server recommended. Using DRBD should ensure the correct read/write condition (but I have small local disk and large NFS share).