I wanted to upgrade my Nextcloud server and host a few other services on my own hardware. The main idea is to use the server to host a bunch of virtual machines to help me run my web design studio and to learn more about container technology such as Docker and LXC.
The server runs Debian GNU/Linux, which is free and open-source software, and I plan to host and play with as many open-source projects as possible.
With the rise of Artificial Intelligence (AI), I also see a huge benefit of having my own server at home where I can experiment with open source AI models like Stable Diffusion without having to rely on cloud computers.
I’ve used Nextcloud on a Raspberry Pi and it worked OK for about a year, but it ran slow and I wasn’t able to use Nextcloud to the fullest.
I also converted an old laptop into a Nextcloud server. The laptop has decent specs to run Nextcloud, but I didn’t want to be restricted by hardware limitations any more.
My homelab is a collection of Raspberry Pi’s, laptops, and older computers that I’ve turned into servers. This setup has worked fine for many years, but the machines were starting to slow me down and needed to be upgraded with faster processors, and more memory and storage.
So it was time to upgrade!
This was my first time using an All-In-One CPU water cooler and I was a bit nervous about the installation.
The kit comes with a couple of USB Hubs where the fans connect together and sync, which is a great idea.
I connected everything but some of the fans didn’t work. After some troubleshooting, I narrow it down to a faulty USB Hub. I ended up opening a ticket with Corsair and waited for the replacement Commander Core Hub.
I noticed that the fans run a little louder now that I have the replacement part, but the noise level is not too bad for having 8 fans.
Proxmox VE is a hypervisor that lets you run virtual machines. So far I have a standalone node with 3 virtual machines and 2 LXC containers.
The best part about my new VM server is the performance. I have all the VM’s running from a 2TB NVMe M.2 drive, so they run really fast with plenty of processors and memory.
I haven’t used the graphics card yet, but I plan to do a GPU pass-through to the VM’s so that I can use the graphics card for AI image generation with Stable Diffusion and Monero (XMR) crypto mining when the GPU is not being used.