New server has been acquired. Debian 13 has been installed.
GS308EP switches have been acquired and installed.
Now, I’m working to migrate to the new machine. 3 1/2 years ago when I started futzing with Docker, I sorta followed guides and guessed, abused it trying to make it do things it wasn’t designed for, and flipped switches I likely shouldn’t have flipped, so the set up is more than a little shabby.
As a result, I’ll likely end more redeploying than migrating the containers.
So rather than go forward with Docker blindly, I want to reassess whether I shouldn’t look into Proxmox, LXC, or Podman instead of Docker, or maybe something else entirely?
Work is just about done dumping ESX for Nutanix, but both of those seem overkill for my needs.
Of course the forums for any of the solutions make their own out to be the best thing since sliced bread and the others useless, so I’m hoping to get a more nuanced answer here.
Don’t go Podman. When I started years ago I installed Fedora with the “containerization” option. This installs podman, not docker as I’m sure most know. I did not.
Podman works great for the most part, but it’s slight differences from docker will have you fighting tooth and nail for certain services to work correctly. And not many (if any at all) have any documentation on getting their containers working with Podman of they don’t start. If you make a GitHub issue asking why or how to get things running in Podman because their docker stack doesn’t work flawlessly like it will in docker, good luck getting help (Mailcow comes to mind specifically here).
Looking back, this decision really shoehorned some very fundamental ideals about containers in my mind, but it was a long fought road I would not choose again. The knowledge I gained about containers with docker would have come soon enough on the easy road.
And yes, you can install Docker on Fedora, but I was much too far down the Podman track before finding out. My environment has changed drastically as of late and most things have been migrated to docker apps in Truenas now, living directly next to their storage as intended (the arr stacks really take a performance hit running their databases over NFS once you have a lot of media for example).
Quick note about Proxmox after coming from ESX myself - it sucks compared to ESX. I’ve tried to move away from it and Nutanix was the closest I could find to ESX, but after my server started complaining it’s drives were not compatible I jumped ship to avoid any write damage to them. I’m downsizing my lab now, I have proxmox running in 3 small NUCs with CEPH storage share and it’s working pretty good. Would love to run ESX or Nutanix instead, but they require a loaf of bread in resource requirements where proxmox only needs a slice of bread in comparison.
What sucks about Proxmox compared to ESX? I’ve never tried it myself
I was a little unfair in my post towards Proxmox. It really is a great solution and I can’t really complain, but it sucks in comparison to ESX where many “custom” items are still hidden in the cli or custom configuration items,. Many of these things are available in the GUI in ESX which is a pretty rough translation for some that have worked in ESX for many years like myself. ESX isn’t without it’s CLI moments but they are rarely ever needed, and if needed only for drastic measures.
The UI is not very intuitive and really looks quite dated too. ESX, Nutanix and XCP-NG have much better interfaces imo, and if Proxmox could throw some of that extra money they’ve earned from the VMware exodus in their UI it would be worthwhile.
Again, I shouldn’t complain but as I get older there’s not much “tinkering” time anymore, and the less time I have to sift through forum posts or official documentation on why something isn’t working as intended, the more easily frustrated I get.