Looking for a recommendation...
I've had to build a new system, put OpenVZ on it, put a CentOS 7 container on it, got plesk installed.
I'm going to test the migration on Monday. If all looks well, I'll bring the system to the data center and do it live.
Since I'm behind a firewall, the public IP's on the container go nowhere. So that's safe. I can keep all the IP's exactly the same (except for the internal addresses).
I have a requirement to keep the public IP's the same as they are in production. Some customers hard-coded their "dedicated" ip's into their apps and systems at offices, so I can't afford to lose customers over that.
I see two approaches to this. I can do this live *twice*, first with temporary IP's, then again to the final container with the originals after the original is taken offline.
Or what would happen if during a down-time I have the same public IP's on both containers and do the transfer/migration? Would that be better? Possibly putting in iptables rules to block the public IP's in VE0? Or temporarily setting the OpenVZ "NEIGHBOUR_DEVS=all" to "detect" or just commenting it out until the migration is finished?
I see the latter as being less work (only one transfer), but using the former would keep things alive but possibly confuse any applications using the static IP's.