🍓 Raspberry Pi Lab - sudo apt update and sudo apt upgrade
🍓 Raspberry Pi Lab -
sudo apt update
and
sudo apt upgrade
#RaspberryPi
Bash
#Linux
Overview
This week I used ChatGPT to build a small maintenance utility for Raspberry Pi OS.
The project's purpose was straightforward: create a Bash script that refreshes package information, upgrades installed software, presents a clean terminal interface, and records everything to a log file.
Not a Tutorial
This article is not a tutorial.
Instead, it's a build report documenting exactly what I built, how I prompted the AI, how I reviewed the generated code, what I changed, and what I learned during the process.
If you decide to build or run something similar, review the generated code carefully and make sure you understand every command before executing it. My goal here is to document my engineering process—not to suggest that anyone should blindly run AI-generated code.
Objective
The project had four simple goals.
- Build a clean Bash utility.
- Display a friendly terminal interface.
- Record execution in a log file.
- Review the generated code before running it.
The Bash itself wasn't the interesting part.
The engineering workflow
was.
Design Goals
The script should:
- Refresh the package database.
- Upgrade installed software.
- Display clear progress messages.
- Stop if a command fails.
- Produce a timestamped log file.
- Remain small enough to understand in one screen of code.
Simple software tends to be easier to review, easier to trust, and easier to maintain.
Prompt #1 - Build the Script
The first prompt was intentionally simple.
Create a Bash script for Raspberry Pi OS that refreshes package information
using sudo apt update, upgrades installed packages using sudo apt upgrade,
displays a clean terminal interface with progress messages, exits on errors,
and writes execution details to a timestamped log file.
Within seconds the AI produced a complete first draft.
That became
Version 1.
Generated Script
#!/bin/bash
set -e
LOGFILE="$HOME/pi-maintenance-$(date +%Y-%m-%d_%H-%M-%S).log"
exec > >(tee -a "$LOGFILE") 2>&1
echo "========================================"
echo " Raspberry Pi Maintenance Utility"
echo "========================================"
echo
echo "Hostname : $(hostname)"
echo "Date : $(date)"
echo
echo "Refreshing package information..."
sudo apt update
echo
echo "Installing available updates..."
sudo apt upgrade -y
echo
echo "Maintenance completed successfully."
echo
echo "Log file saved to:"
echo "$LOGFILE"
Even before running the script, there was plenty to evaluate.
Prompt #2 - Review the Script
Rather than immediately executing the code, I asked the AI to perform a second job.
Review this Bash script as though you are performing a professional Linux
code review. Identify safety concerns, security issues, edge cases,
error handling, logging improvements, portability concerns,
maintainability issues, and opportunities to simplify the design.
Do not rewrite the script unless absolutely necessary.
Explain your reasoning.
I find this second prompt at least as valuable as the first.
Generating software is becoming routine.
Evaluating software
remains an engineering discipline.
Review Findings
The review identified several discussion points.
| Area | Observation |
|---|---|
| Overall design | Clean and readable |
| Safety | Suitable for manual execution |
| Logging | Good foundation; could include more system details |
| Error handling |
set -e
prevents continuing after failures
|
| User interaction | Could ask for confirmation before upgrading |
| Maintainability | Short, understandable, easy to extend |
| Future enhancements | Disk checks, reboot detection, colored output, dry-run mode |
None of these observations prevented the script from running.
Several, however, made a future Version 2 an obvious improvement over Version 1.
Human Review
After reading the AI's assessment, I performed my own review.
Questions I asked included:
- Do I understand every command?
- Is anything happening automatically that shouldn't?
-
Would I be comfortable scheduling this script with
cronlater? - Does the log contain enough information for troubleshooting?
- Could another Raspberry Pi owner understand what this script is doing?
Only after answering those questions did I decide to execute it.
Running the Script
The script was saved locally, made executable, and launched manually.
$ chmod +x pi-maintenance.sh
$ ./pi-maintenance.sh
========================================
Raspberry Pi Maintenance Utility
========================================
Hostname : raspberrypi
Date : Sat Jul 25 09:14:37 CDT 2026
Refreshing package information...
Hit:1 http://archive.raspberrypi.com/debian bookworm InRelease
Hit:2 http://deb.debian.org/debian bookworm InRelease
Reading package lists...
Building dependency tree...
Reading state information...
Installing available updates...
12 packages upgraded.
Maintenance completed successfully.
Log file saved to:
/home/pi/pi-maintenance-2026-07-25_09-14-37.log
The important point is that
I
chose to execute the script.
The AI generated the code.
The AI reviewed the code.
The human accepted responsibility
for running the code.
That feels like a healthy engineering
workflow.
Sample Log File
One of my favorite additions was automatic logging.
========================================
Pi Maintenance Utility
========================================
Run Started:
2026-07-25 09:14:37
Hostname:
raspberrypi
User:
pi
Updating package information...
SUCCESS
Installing available updates...
SUCCESS
Packages upgraded:
12
Run completed successfully.
Elapsed time:
1 minute 43 seconds
========================================
A log file transforms a simple utility into something much easier to troubleshoot and improve over time.
Testing Environment
Hardware:
- Raspberry Pi 5
- 8 GB RAM
Operating System:
- Raspberry Pi OS Bookworm
Execution:
- Manual terminal launch
Status:
- Successful
Lessons Learned
The biggest surprise wasn't that AI could write Bash.
That part is
becoming expected.
The more interesting discovery was how valuable AI
became as a reviewer.
One prompt generated the code.
A second
prompt critiqued the code.
Finally, a human reviewed both before anything
was executed.
That sequence resembles a traditional software workflow more than the popular image of "vibe coding."
The result wasn't simply a script.
It was a documented engineering
process.
Ideas for Version 2
Possible improvements include:
- Colored terminal output
- Internet connectivity test
- Available disk space check
- Reboot-required detection
- Optional confirmation before upgrading
- JSON log output
- Email notification after completion
- Dry-run mode
- Automatic log rotation
Each enhancement is small on its own.
Together, they turn a basic maintenance script into a more polished system utility.
Final Thoughts
There's no shortage of AI-generated code on the Internet.
What's still
relatively uncommon is documenting everything that happens
after
the code appears.
That's the purpose of these Build Reports.
The prompts matter.
The review matters.
The testing matters.
The
log files matter.
Most of all, the engineering judgment matters.
The AI wrote a useful first draft.
The review improved it.
The human made the final decision.
That's a workflow I expect to use
again.
Build Information
Project:
Pi Maintenance Utility
Category:
System Administration
Language:
Bash
Platform:
Raspberry Pi OS
AI Assistant:
ChatGPT
Development Approach:
Prompt → Generate → Review → Refine → Execute → Document
Estimated Development Time:
Approximately 20 minutes
Status:
Working
Difficulty:
Beginner
Tech Reader Labs:
Build Report #001
Aaron Rose is a software engineer and technology writer at tech-reader.blog.
Catch up on the latest explainer videos, podcasts, and industry discussions below.
.jpeg)
