Well, right now I am happy, but I recently changed my setup from a complex Debian with open-zfs storage to a simpler and more up to date Fedora-server on btrfs version and while btrfs is not nearly as advanced as open-zfs it is much less of a hassle to maintain and less likely to break on Linux kernel updates.

Man I just need to set up a proper SAN, having a bunch of servers without any ability to link storage togetherbetween them gets kind of annoying

Also network seperation…

I have a small home lab, best thing I ever did was put it all in containers. Rebuilding is as simple as copying my docker data folder and starting them back up

I would have used Docker and Docker Compose a whole lot more. There’s servers I’ve realized I’ve been hosting for over a year and forgotten about. At least with Docker I can check on those much more easily

Note: I use hyperV, so my views here may not be applicable for all.

But, get a server-grade motherboard from the start. Supermicro is cheap used.

Having IPMI / out of band management is incredibly helpful to troubleshoot issues.

Stability seems to be a lot better.

Drivers are more reliable on WinServer.

My homelab started as an experiment when I discovered docker and proxmox. I spun up things randomly without paying much attention to resource allocation and any form of order. I ended up using a lot of these services in my daily workflow and they became central to my daily workflow. Now, I am running low on resources on some of my servers due to over allocation and have docker containers randomly in different VMs. I wish I thought this journey through and realized that I will be doing this homelabbing for the long run.