All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Any novel tricks for upgrading physical CentOS 6 servers to AlmaLinux 8 servers?
LosPollosHermanos
Member
These are in datacenters I don't have physical access to. For simplicity I am just asking about vanilla servers. Oh, and it also needs to be done without significant downtime. 
Some will say it's impossible but I'm not convinced and love a challenge. They all use SW RAID 1. I was thinking that maybe I could degrade the array by removing one of the drives, then clone over the server from another server I have preconfigured onto the disconnected drive using something like Clonezilla, then I reboot onto the cloned drive. Once that is confirmed working in a degraded state I can format and re-attach the former RAID drive. If something goes wrong before that I can still boot back into the old degraded RAID drive. I will of course have everything backed up before I do this.
The first thing a lot of people will say is to just set up new servers. It's not really an option for many of them for various reasons. I don't want to go off on a tangent trying to explain why, it just is. I've done really difficult things on Linux before so something like this doesn't scare me too much. I would also be testing this on non-production servers beforehand.
Comments
I hope you understand that you can't just run a single command to convert an OpenVZ VPS to a KVM.
Why not to get a new server as node setup with new VMs and migrate all the data and migrate all the IPs to new node - almost no downtime if it’s done correctly.
I've done it many times in both directions.
Read the last paragraph of my original post. I edited it about 5 minutes ago so maybe you didn't see it.
I rewrote my original post to just talk about vanilla servers to keep it simple.
So many red flags...
:facepalm:
To those not being silly, they'd know to backup the data, reinstall Rocky 8 and then restore data in far less time.
You already have the data being backed up, so nothing extra needed.
If the most straightforward strategy doesn't work for you, you're doing it wrong.
Like what? How else am I going to upgrade without downtime? What I could probably do is get another server with similar hardware in the same datacenter and set that up the same way, mirror the VMs over and confirm everything works, then do the same thing on the original server and move everything back. Still lots of downtime but I could transfer the most important customers first to mimimize it for some of them.
I did find this ansible script which I might be able to modify to do what I want. Even if I could just get to CE7 that would be ok. I could use Elevate from there.
https://github.com/godaddy/centos7-upgrade-scripts/blob/master/README.md
https://linuxiac.com/almalinuxs-elevate/