First Experimental Home TrueNAS Build on Low End Hardware: Core or Scale?

Hello,

I’ve watched Tom’s videos and read a lot, and I’m still going back and forth over whether I should go for Core or Scale in January 2024 for a low end 4 bay home storage server.

For the longest time, I’d understood that Core was preferred if the goal is just to be a storage server, and Scale was best for added features (virtualization, apps, etc.), but lately I’ve read/heard more Reddit Storage Warriors say Scale is best as it’s more stable under certain conditions (not sure how Linux is more stable than BSD, but…).

So now, I’m indecisive again. I would appreciate some perspective and advice.

Details:

I’m setting up a low end Dell Optiplex 5040 (i7-6700 4c8t, 32 GB DDR3L, 2x10GbE SFP+ card, LSI HBA) as my first ever TrueNAS build, both to use for actual storage here on my home network and also to learn to use TrueNAS before investing in more powerful hardware and moving away from my current QNAP solution entirely.

For this little server, I have 4x1.9 TB 2.5" SSDs available for storage. I’m thinking of setting them up in two mirrors (one for live VM drive image storage in Proxmox and the other for in-LAN data that will benefit from fast, silent storage (database server, essential backups, etc). Maybe a striped mirror?

(I’d do a 4 disk RAIDZ1, but I understand that is a terrible idea for VM storage.)

I wouldn’t mind running a few microservices on the Dell if it has the horsepower to do so, but I also don’t know a thing about Kubernetes and would rather not use Docker if I can avoid it. I’m a big fan of LXCs in Proxmox.

I’m also still learning how to really use Proxmox and do things like GPU passthrough and the like. One thing that attracted me to Core is that it’s really just a NAS and it would be easier to learn alongside learning Proxmox than Scale, with its integrated virtualization features that are … not Proxmox.

The most important thing is don’t get paralyzed by the choice. They’re both great products, so you won’t be losing out either way.

I’ve got a couple Core systems running and I’m very happy with them, but I could reinstall, import my ZFS pool and I’d be just as happy on Scale.

1 Like

Thanks!

Choice paralysis is definitely something I’m struggling with a bit.

My biggest hesitation with Core is that I’m not nearly as good on a BSD command line as I am in Linux

But, realistically, TrueNAS Core is an appliance OS.
How often am I really going to need to dive under the hood?

You don’t really need the command list for daily use in Core or Scale and BSD is not that much different for the command line. If it’s just storage, Core is fine.

1 Like

Thanks for taking the time to read my post–I’m sure I’m the 52 millionth person to ask. :slight_smile:

I’m going with Core. Installing it now, actually. (Pondering whether I need a 16GiB swap partition on my boot NVME. Probably doesn’t hurt anything…)

I think one reason I hesitated was that I have more trouble getting system diagnostics info when I’m stuck in a terminal on BSD. There’s no lm-sensors package, or watch command, or other little things I’m used to using for system monitoring. I had to rescue a broken OPNSense install once and didn’t realize until I started how quickly all the little differences between Linux and BSD can pile up when you’re in the CLI weeds. :slight_smile:

TrueNAS should expose all that sort of diagnostic data in the Web GUI for me, I suspect.

The Core installer is asking me if I want to set up a 16 GiB swap on the boot disk (500 GB Gen 3x4 NVME SSD). I have 32 GB of DDR3-1600 in here, and there’s more than enough room on the boot drive, but I’m not sure what the preferred way to go here is.

I never want it to need swap, and I doubt I’ll ever actually run out of memory (it’s basically just me using this thing, with ~4 TB array for VM storage and other quick/silent access things).

Is there some reason it needs to be on?
If I expect never to need it, will enabling it cause other potential drawbacks?

I’d say there’s no downside to having swap available unless you’re severely space constrained. Even if it’s unlikely, running out of memory can cause some real problems, so better safe than sorry, I think.

Also, as far as keeping tabs on sensors and stuff like that, consider setting up netdata on the machine.

1 Like

Thanks for the rec re: netdata. I remember Tom doing a video on that a while back.

I was looking at deploying CheckMK, as I like its Proxmox integration, so I’ll have to see how TrueNAS Core plays with that, as well.