Howto make an old PC function as a Home NAS

[rsnfunky]

Hi,

As the question describes, the aim is to get an old PC function as a home NAS to which the following can be connected:

1) TVs for playing media saved on HDDs
2) Mobile Phones
3) Linux / Windows PCs / Laptops to access and save data
4) If possible access this data from net using a VPN

I was reading this article: https://www.xda-developers.com/how-turn-old-pc-into-nas/

This was on TrueNAS and I understand that it requires one dedicated disk to boot.

My Aim:
1) Have the same features like TrueNAS but using a gentoo box. Can also make it dual boot with Windows and connect to a TV (Functions as a PC in another room).

2) Install Softwares like:
a) Nextcloud for working like Google Photos and backup
b) Jellyfin for Media Server - Will this also fuction as a DLNA for playing media on home smart TV?
c) Any google drive alternative for auto back android devices and whatsapp backups
d) Any other softwares which can be recommended?
d) How to get secure access from the internet.

3) I can connect to this old PC NAS, all the internal and external HDDs that I have and these will be availabe whenever I am at home and that too from any device on the wifi.

[szatox]

Sounds like a plan.
What's the actual problem though? You're unlikely to get a full-blown custom guide, just ask specific questions once you get stuck on something.

[lars_the_bear]

I would imagine that any Linux distribution would be capable of providing basic NAS functions. It really depends on the amount of set-up you want to do. I don't think using Gentoo here would offer an advantage over any other distribution -- there would still be a lot of set-up. I guess that's where platforms like TrueNAS score: they'll do most of the basic stuff with little additional set-up. My understanding is that TrueNAS can only be installed on a complete drive, but storage is cheap, and you could still dual-boot by using the PC's BIOS to select a drive to boot from (rather that, say, using multiple partitions on the same disk).

But I fear that you might soon get fed up with the constant fan noise and poor energy-efficiency. In my view, it's only worth using PC-type hardware for NAS if you plan to run more CPU-intensive stuff on it as well. Otherwise it's a waste of resources. Just my $0.02, of course.

BR, Lars.

[NeddySeagoon]

Build on what you know works. Add thigs so you know what breaks. :)
and can back it out.

Start with NFS as that's fairly simple to make work.
Add Samba for the Windows systems.

Do not install everything you think you need and tell us it's not working.
We already know that. :)
The idea is to keep the problem space small, so it's easy to diagnose and fix.

Actually the order you add bits is not important. It's the one at a time that matters.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.

[Bob P]

Using an old PC for a NAS is a great idea -- but if you let feature creep determine how much functionality you want, you're likely to reach the point where old hardware may not fit the bill and new hardware will be required.

I've been doing this for a long long time. The demands that we placed on old PCs that were re-tasked as home servers were a lot simpler back then. My first "NAS" was built long before the term "NAS" became a marketing thing. Back then I just ran linux on an old PC and installed SAMBA to serve Win and Linux clients at home. Back then we just called them SAMBA servers. The hardware requirements for that sort of thing in the home were so minimal that you could literally use any old boatanchor and a few big IDE disks to do the job. Way back when I used a 486 and then Pentium box for this sort of thing. Ultimately the disk and networking interface requirements forced motherboard/CPU upgrades.

I was deep into Gentoo at the time, yet for the NAS I moved to a FreeNAS to do file serving, backup, and media streaming. It was just simpler to use a turnkey distribution than to try to build everything by hand. Back then just building a LAMP server on Genoo was a major project and a full blown multi-feature NAS was beyond the reach of most people.

If I were in your shoes, I wouldn't start off trying to do it all under Gentoo -- there will just be too much heavy lifting while you're learning the ropes. If it were me, I'd start off with a turnkey distribution, get a basic system up and running, and then add features as needed, taking incrementally small steps. If you really want to do it in Gentoo, I'd recommend building the Gentoo box after the turnkey box is already up and running. I say this because getting your full featured NAS up and running as quickly as possible will be a lot more rewarding than the delayed gratification that will come from a full Gentoo build.


Regarding the dual boot idea, I can't really say that I would ever consider building a NAS and then taking it offline to boot windows on that metal. It just seems counterproductive to me to take down the NAS for any secondary task, but that's just me.


With the older version of TrueNAS (FreeNAS), the dedicated boot device could be a USB stick. Back in my FreeNAS days (prior to the advent of SSD) I used mirrored USB sticks as boot devices for fault tolerance. That was a great solution as it allowed me to reserve the machine's SATA ports for the data disk array. When they rebuilt FreeNAS and the devs were using SSD, they lost USB compatibility and forced everyone to use SSD for the OS mirror. The result was that they consumed 2 SATA slots that put limits on the number of disks in the zpool. Everyone ran out of SATA ports and ended up migrating to LSI SAS disk controllers as the disk arrays grew and grew.

The moral of that little story is that things get very complicated very fast, and the more features you want, the more you're going to strain your hardware, and the more likely that the prospect of building a very complex modern system out of old hardware is likely to fail. You're likely to end up being forced into new hardware. My suggestion would be to start off simple, and add features incrementally, so you'll know what features force you into upgrades and you can make decisions about whether or not they are worthwhile and fit the budget.

Neddy's recommendation to keep the problem space small, and perform incremental feature expansions are true words of wisdom.