DG
Back

Building a NAS with Proxmox and Mergerfs

#programming #linux #homelab #computers

Updated 27 Jan 2025

Today I put together a NAS that I hope to use for a while. For now, it will be my main storage target but hopefully soon, it will become my backup. I have two cases but only one CPU/motherboard. I want to reserve the bigger, better case for a better system so for now, this mostly free, slightly janky system will do. It's a tight fit with the motherboard, power supply and drives so I'm avoiding HDDs altogether in this build. I don't have a lot of data (yet) so I'm only installing two drives for data — a 2TB Crucial BX500 SSD and a 1TB Crucial P3 SSD that I'll use for Time Machine backups.

The nice thing about this setup is that I can add more SSDs as my needs grow since I'll be using mergerfs which allows me to make multiple mismatched sized drives appear as one single volume. It's a union filesystem that supports addition and removal of drives with no rebuild times. It doesn't come with parity so I'll also be using SnapRAID for parity once I have more drives.

My apologies if the rest of this post is a little dull — it's largely going to be documentation for what I did so that I can refer back to it if needed. It's largely based on Alex Kretzschmar's awesome guide to setting up a media server with Proxmox as the host.

Proxmox

I installed Proxmox on a 128GB NVMe drive. After rebooting, I ran the Proxmox VE Post Install script from the awesome Proxmox helper scripts repo. It provides options for managing Proxmox VE repositories, including disabling the Enterprise Repo, adding or correcting PVE sources, enabling the No-Subscription Repo, adding the test Repo, disabling the subscription nag, updating Proxmox VE, and rebooting the system.

bash -c "$(wget -qLO - https://github.com/community-scripts/ProxmoxVE/raw/main/misc/post-pve-install.sh)"

The version of mergerfs from the Debian upstream repos significantly lags behind the most recent release available on Github so I opted to use the command below which will query the Github API for the latest release, query my local OS for which release I'm running and then download the correct .deb package before installing it with dpkg.

curl -s https://api.github.com/repos/trapexit/mergerfs/releases/latest | grep "browser_download_url.*$(grep VERSION_CODENAME /etc/os-release | cut -d= -f2)_$(dpkg --print-architecture).deb\"" | cut -d '"' -f 4 | wget -qi - && sudo dpkg -i mergerfs_*$(grep VERSION_CODENAME /etc/os-release | cut -d= -f2)_$(dpkg --print-architecture).deb && rm mergerfs_*.deb

After it ran successfully, I set it up to run on a cron job every few months. Otherwise the package will never update since I'm not using the repo version.

Storage

Then it was time to get my main data drive set up. I ran inxi -xD to make sure it was showing up. Then I used gdisk to create one large partition spanning the whole drive and mkfs.ext4 /dev/sda1 to create an ext4 filesystem. Then I created a mountpoint for my data disk as well as a mountpoint for my mergerfs pool:

mkdir /mtn/disk1
mkdir /mnt/storage

Lastly, I added entires to my /etc/fstab file for my data disk and mergerfs pool:

# <file system> <mount point> <type> <options> <dump> <pass>
/dev/disk/by-id/ata-CT2000BX500SSD1_2427E8BA6593-part1 /mnt/disk1 ext4 defaults 0 0
/mnt/disk* /mnt/storage fuse.mergerfs defaults,nonempty,allow_other,use_ino,cache.files=off,moveonenospc=true,dropcacheonclose=true,minfreespace=200G,fsname=mergerfs 0 0

The options in that mergerfs fstab entry just set some sane defaults. I ran mount -a to reload the fstab entries and df -h to verify the mountpoints:

root@rackchoice:~# df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/sda1             1.8T   28K  1.7T   1% /mnt/disk1
mergerfs              1.8T   28K  1.7T   1% /mnt/storage

Docker, Samba, and NFS

I installed Docker by running curl -fsSL https://get.docker.com | sh and Samba by running apt install samba. I then created a file at /etc/samba/smb.conf that looked like this:

[global]
    workgroup = DCG
    server string = rackchoice_samba
    security = user
    guest ok = yes
    map to guest = Bad Password
    log file = /var/log/samba/%m.log
    max log size = 50
    printcap name = /dev/null
    load printers = no

# Samba Shares
[storage]
    comment = Primary Storage
    path = /mnt/storage
    browseable = yes
    read only = no
    guest ok = yes

[TimeMachine] 
	comment = TimeMachine backups
	path = /mnt/timemachine 
	valid users = timemachine 
	writeable = yes 
	vfs objects = catia fruit streams_xattr 
	fruit:time machine = yes 
	browseable = yes

At this point, I realized two things:

  1. I hadn't mounted my second NVMe drive in my system that I was using for Timemachine backups
  2. I had been doing everything as root and I needed to create a dominic user and a timemachine user.

So I ran the following commands:

sudo mkdir -p /mnt/timemachine
sudo chmod -R 777 /mnt/timemachine
adduser timemachine
smbpasswd -a timemachine
smbpasswd -e timemachine
mount /dev/nvme1n1p1 /mnt/timemachine/
adduser dominic
smbpasswd -a dominic
smbpasswd -e dominic
systemctl restart smbd

I also added the timemachine mountpoint to my /etc/fstab.

Conclusion

And that's it! The NAS is up and running. I connected to the SMB share with my Mac as well as a Raspberry Pi on my network. The next step is to get my media files transferred over and then deploy Jellyfin on my k3s cluster (with FluxCD) and point it at the Samba share.