Home · Volume 2 · Chapter 01

📖 Chapter 01 — The Family as a System

The NAS isn't for you. It's for your family. That changes everything about how you design it.

v0.1 · draft Vol 2 · Ch 01
~8 min

Learning Objectives

Introduction

You built the NAS. You know what a vdev is. You configured the snapshot schedule. You set up RAIDZ2 with 6 drives. The technical foundation is solid.

Now the harder problem: getting the rest of the family to actually use it.

This chapter is short but it sets the tone for the rest of Volume 2. The technical choices in Volume 1 (datasets, permissions, snapshots) were about what works. The choices in Volume 2 (apps, layouts, defaults) are about what people will use. Those are different questions.

The principle of least surprise

The most useful design principle for a family NAS is also the simplest: do what people expect.

When Mimi opens File Explorer, she expects to see the family photos in \\NAS\Photos. When she clicks the family photos folder, she expects to see folders organized by year or event. When she opens a folder named "Hawaii 2024," she expects to see the photos from that trip — not a ZFS dataset, not a sync conflict, not a half-uploaded file from her phone.

Every deviation from what people expect is friction. Friction is what makes people stop using the system and go back to iCloud or Google Drive. The job of Volume 2 is to remove friction.

Who is "the family"?

For TK's build, the family is four accounts (tk, Mimi, Lai) with very different relationships to the technology:

When you design the family experience, you're designing for Mimi. TK can handle complexity. Mimi will just leave.

Design heuristic

For every decision, ask: "Would Mimi understand this without me explaining it?" If the answer is no, simplify. The next time TK is on a flight, the family needs to be able to use the NAS without him.

Three failure modes

Three ways a family NAS fails, all of which come from designing for the technical user instead of the family:

1. The "where do I put this?" problem

You set up a beautiful dataset tree with 15 carefully-named datasets. The family uses it for three days, then someone saves a file to their Desktop and forgets about it. After a month, half the family's data is on a laptop hard drive that wasn't backed up.

The fix: make the right place obvious. When Mimi plugs in her phone, the photos should go to Immich automatically. When she downloads a file from email, it should land somewhere that gets backed up. The default behaviors should be the right ones.

2. The "I can't find it" problem

You set up Immich with face recognition, OCR, location tags, albums, and search. The family uploads 5000 photos. Nobody can find the one they want because they don't know which app to use, which search to do, or even that the photo is in Immich.

The fix: make the obvious path the only path. If the family has a single way to access photos (Immich on the phone, the Immich web UI on the laptop), and you train them to use that, search will work because they're searching the right index. Don't give them five ways to access the same data; give them one good way.

3. The "it broke" problem

TrueNAS updates itself. An app updates itself. A drive reports a SMART warning. The system has a power blip. Now the family can't watch their movie.

The fix: make recovery automatic and invisible. Most issues should self-heal (services restart, snapshots roll back, scrubs repair). For the issues that can't self-heal, the family should know how to reach you. The support runbook ("if X happens, do Y; if Y doesn't work, message TK") should be short and live somewhere obvious.

The "Mimi test"

For every change you make, run the "Mimi test":

  1. Set up the change.
  2. Pretend you're Mimi — non-technical, busy, doesn't want to learn.
  3. Try to do the thing the change is supposed to enable.
  4. If you can't do it without instructions, redesign until you can.

Examples of the Mimi test in action:

The hard parts of family IT

The conversation was honest about what's hard:

None of these are problems you solve with technology. They're problems you solve with attention, communication, and being the person who notices when something's wrong before anyone else does.

Engineering Note

The hardest part of running infrastructure isn't the configuration. It's the human factors. The 6-drive RAIDZ2 with monthly scrubs and weekly SMART tests is the easy part. Getting the family to actually use the system — and to keep using it for years — is the hard part.

The good news: most of the work is done by being thoughtful upfront. The rest is being available when something goes wrong. Neither requires a network engineering background; both require showing up.

Summary

The NAS is a family system, not a personal project. Design for the least technical user in the house. Use the principle of least surprise. Make the right behavior the default behavior. Make recovery automatic. Be available. None of this is technical; all of it is what makes a home lab survive past the first six months.

Checklist

Looking Ahead

Chapter 02 is the chapter on photos — the highest-value data, the app the family will use most, the one with the most "wow per minute of setup." Immich in production: setup, face recognition, albums, sharing, the migration off Google Photos.

Ch 01 · v0.1 · drafted from the original ChatGPT conversation, July 2026