My company currently has a RHEL 7.8 server running 4 dockers that have Ghost installed in each and each instance hosts external facing corporate websites (static content, updated by an internal content developer only). We identified an issue with the configuration of the Dockers and need to reconfigure the installations (read: delete and rebuild), but would really like to avoid the downtime of taking the sites offline. It had been our plan to stand up a second server with the same configurations, but with the corrections in place, and maintain a two server set up to allow for taking the primary server offline to make the corrections, hopefully preventing that downtime window. We aren’t worried about performance under load, but about availability. We have this configuration because it was previously indicated that we wouldn’t be able to host multiple URLs in a single instance of Ghost, so each instance was set up in a Docker to isolate it and allow us to install them all on a single server, rather than requiring multiple servers for the sites.
The only way that would work is if you spun up the second server after the first is taken offline. Changes on one instance will not propagate to any other running instance which is why we explicitly state Ghost doesn’t support any multi-server setups.
If you want to avoid downtime for visitors you’re better off setting up a cache that will serve cached content when the upstream isn’t available. Admin and some members functionality won’t be available whilst switching but the site itself won’t be offline.
So we aren’t looking to have changes propagate to each other- these would be updated individually by the content owner and not synchronized between the servers, if that makes any sort of difference. (I’m guessing not, though)