The Real Cost of a DIY NAS
(updated )
A year after building my own NAS, I wanted to revisit the subject of when you should build a DIY NAS or buy a prebuilt one. I’ve had enough time to experience the hidden costs of building my own NAS and for my opinions to mature.
Troubleshooting Problems
Buying a prebuilt model means the manufacturer is a single point of contact for problems. Since I built a DIY NAS instead, this meant that I was on my own to solve any problems that arose. In the past year, I’ve had to deal with the following issues:
A rare ZFS bug causing permanent metadata corruption. The only workaround was recreating the ZFS pool from scratch, which is incredibly inconvenient and causes downtime. I’m still running ZFS and am hoping to avoid triggering it again since the bug is still unresolved.
To be fair, if I encountered this bug on a prebuilt NAS running ZFS, I doubt the outcome would have been any different since this is a ZFS issue.
Crashes with the Intel I225-V controller on the Asus motherboard’s onboard ethernet port, which permanently breaks network connectivity until I restart the NAS. A nightmarish experience with Asus support bricked the onboard ethernet controller. I had to buy a replacement NIC to fix this issue.
A Linux kernel regression in the drivers for the replacement NIC causes the same network instability problems that I experienced with the motherboard’s onboard ethernet port. I’ve worked around the issue temporarily by downgrading to a previous version of the kernel, but I’ll have to find a better solution if the drivers aren’t fixed. I seem to be a magnet for networking problems 😢
I have spent dozens of hours on these problems, and more than a few stressful evenings and weekends.
I am documenting my experiences to provide a realistic picture of the sorts of troubleshooting that you may be in for. If you are comfortable with the potential time and effort and have the technical knowhow, don’t let this dissuade you from building your own NAS.
Otherwise, getting a prebuilt NAS would likely have prevented the networking problems, since NAS manufacturers offer a limited set of models and can ensure software and hardware compatibility. You may not be getting the latest and greatest Linux kernel with a prebuilt NAS’s operating system, but the likelihood of experiencing issues is also lower.
NAS or Tinkering Platform?
If you only need NAS features like file sharing, backups and running basic applications, I recommend a prebuilt unit. After my experiences, I would be happy to spend a bit more money for a prebuilt model to save time spent on troubleshooting. Prebuilt NAS operating systems limit the changes that can be made, reducing the chances of something going wrong. This was how my previous NAS, a Synology DS918+, ran flawlessly for 5 years until it died.
However, that recommendation comes with a significant caveat. Choosing a prebuilt NAS means buying into the manufacturer’s ecosystem, putting you at their mercy for the long-term viability of your setup. The single point of contact for problems is also a single point of failure for support and features. For example, Synology is now locking their newer NAS models to overpriced Synology branded drives, and surprising users with removal of features between major operating system updates.
These risks, on top of my previous Synology NAS being out of warranty pushed me to the DIY route. I also wanted a lot more out of my machine. I needed a platform for tinkering and experimentation. My machine runs Proxmox, which allows me to run virtual machines and LXC containers. NAS features are provided by an LXC container that serves files over the network.
With a DIY build, I own the problems, but I also own the solutions and the freedom to evolve the system as I see fit.
Conclusion
The true cost of building a DIY NAS extends far beyond the parts list. While I spent S$1169.5 on hardware, I’ve invested countless hours troubleshooting ZFS corruption and networking issues. These aren’t one-time setup costs - they are ongoing responsibilities that come with being my own system administrator.
Despite these challenges, I don’t regret the decision. My DIY build gave me a powerful, flexible platform that has evolved from a simple NAS into a home server supporting virtualization and experimentation.
The choice comes down to which set of risks you’re willing to accept. If you need the reliability of an appliance for basic NAS functions, a prebuilt unit is a great starting point, as long as you accept the risks of lock-in and a closed ecosystem. However, if you want an incredibly rewarding and powerful platform for learning, the DIY route is unmatched - provided you don’t mind becoming your own on-call support for those occasional 2AM troubleshooting sessions.