When a Discourse upgrade fails due to low disk space, the UI sadly breaks with a 500 error. I then have to run ./launcher rebuild app from the terminal, which downloads all Docker images first, pauses to let me know there’s less than 5GB of space available, waits for me to press ‘y’, and then takes some time to cleanup the old images. Then, I have to run the rebuild command again, which starts by re-downloading all the Docker images.
This process has many steps, some of which need manually checking the logs, and they all add up to more downtime for the forum. Luckily, this issue doesn’t happen often to me. Though, I still get nervous when the forum is down ![]()
This downtime looks entirely preventable to me. Here’s my proposed enhancement. Before upgrading, the web UI can prompt the user:
You have less than 10 GB of free disk space. Would you like to clean up old containers before upgrading?
If the user accepts it, then the disk is cleared first, and they are informed:
Cleared 10GB of old containers. You now have 15GB of free disk space. Feel free to start the upgrade now.
And now they can confidently run the Discourse upgrade from the Web UI ![]()