Discourse herstellen na volle schijf

AI-generated summary

User dhoop experienced a Discourse outage after their 16GB AWS EC2 instance ran out of disk space. After freeing space via rm -rf /var/lib/docker/overlay2 and journalctl --vacuum-time=10d, they rebooted but encountered “Connection refused”. Attempts to rebuild with ./launcher rebuild app failed due to insufficient disk space (<5GB). IAmGav suggested ./launcher cleanup, while Stephen noted 25GB is the recommended minimum to avoid recurrence. Ed_S shared that 20GB can suffice for small forums if backups are managed, and noted journal logs can be trimmed with journalctl --disk-usage. dhoop later expanded the disk to 40GB (29% used) but still faced issues starting Discourse, with launcher rebuild app failing due to missing image local_discourse/app:latest. The error log is here. dhoop temporarily fixed the rebuild by switching postgres.template.yml to postgres.10.template.yml in app.yml, but is unsure whether to keep this change. Ed_S advised consulting the PostgreSQL 12 update thread for support, as the issue may stem from a failed PostgreSQL upgrade.

My AWS ec2 instance ran out of disk space (it has 16GB). I ran rm -rf /var/lib/docker/overlay2 and journalctl --vacuum-time=10d to free up space, and now df says I have 7GB of free space. after rebooting the instance I get Connection refused when trying to connect to the page.

I tried sudo ./launcher rebuild app but it fails saying there less than 5GB of disk space.

Any suggestions?

try

./launcher cleanup

25GB is usually the recommended minimum. Even if you can get it back up running with your current allocation this is going to happen again with such a small volume.

Just to note, Hetzner’s lowest tier offers only 20GB, but that’s been working for me. It can be worth tidying up the backups directory before trying an upgrade: having, of course, taken a new backup and copied it elsewhere for safety. (It’s a small forum: a backup is 120MB.)

In another case, I use Digital Ocean’s lowest tier, offering 25G, and in that case the forum’s backup comes in at 640MB.)

The one time I had trouble, it was the journal thing, but I cut that right down:

# journalctl --disk-usage
Archived and active journals take up 104.0M in the file system.

Ok, I have increased the size of the disk so discourse-doctor sees 40G (29% used) so space is no longer the issue. I need help getting discourse up and running.

sudo ./launcher start app fails with
Unable to find image 'local_discourse/app:latest' locally

The failing output of sudo launcher rebuild app is here

How do I fix this?

ok I got sudo launcher rebuild app to work by changing “templates/postgres.template.yml” to “templates/postgres.10.template.yml” in app.yml

Should I leave that change, or change it back and try rebuilding?

I think you need to visit the postgres update thread and ask for help…

1. UPGRADE OF POSTGRES FAILED

 * Please visit https://meta.discourse.org/t/postgresql-12-update/151236 for support