Sander

ℹ️ accounting January 2026 ℹ️

People have been generous donating in the first year of month for mountains.community. I do realize, that some donations are meant for the whole year. but that sends a signal of trust (and wanting to keep hanging around).

Thanks all, you rock.

Full history: https://mountains.community/behind-the-scenes#accounting

#MountainsAdmin

Feb 14 at 11:25amWeb
Show threadMark Pospesel Feb 14
@bergmeister thank you for all your work! 🙏
Show threadmarcoFeb 17
@bergmeister It's crazy what these servers cost. So the S3 buckets are 'only' 1.5TB total? I'd assume that they'd be full pretty quickly?
Show threadSanderFeb 17

@marco housekeeping does the trick (media that comes in from remote instances and is not interacted with during the last 4 days are removed, will be refetched when needed).

The S3 media bucket for mountains.social is hovering between 160 and 200 GB (the peak you see is because of AI scrapers that I then blocked). That is hosted on UpCloud.

The S3 backup bucket (at Hetzner) is roughly the same size. I do not keep media backups in the classical sense (for the VM and database I do), where I can go back X number of days. I do sync the media once a day, so in case of catastrophic failure of the media bucket, I do have a recent copy.

Show threadmarcoFeb 17
@bergmeister that's very neat! Is that a server-wide config or administration setting? My private instance, that only I am on, is already at ~60GB after three months
Show threadSanderFeb 17

@marco 60 GB is indeed a lot for a single user instance.

These are the commands I run as part of a script every day (I forgot I lowered the 4 days to 3 to cope with the AI scraping).

----------------------
echo "Processing: tootctl media remove"
RAILS_ENV=production NUM_DAYS=2 bin/tootctl media remove --days=3

echo "Processing: tootctl preview_cards remove"
RAILS_ENV=production NUM_DAYS=30 bin/tootctl preview_cards remove --days=30

echo "Processing: tootctl media remove --remove_headers true"
RAILS_ENV=production bin/tootctl media remove --remove_headers true

echo "Processing: tootctl media remove --prune_profiles true"
RAILS_ENV=production bin/tootctl media remove --prune_profiles true
----------------------

You can add "--dry-run" on a lot of tootctl commands if you just want to try it out without actually changing stuff.

Check out the documentation here: https://docs.joinmastodon.org/admin/tootctl/#media-remove

Using the admin CLI - Mastodon documentation

tootctl commands that can be run from the CLI.

Show threadmarcoFeb 17
@bergmeister Thank you for this! I'll definitely use this! :)
Show threadSanderFeb 17
@marco I would be interested to know how much you were able to save on S3 storage afterwards.
Show threadSanderFeb 17

@marco by the way, with

RAILS_ENV=production NUM_DAYS=2 bin/tootctl media usage

you can check the current usage across the various objects and therefore see where you were able to save.

Show threadmarcoFeb 17
@bergmeister I'll let you know once I get around to implementing this!
Show threadmarcoFeb 28
@bergmeister that's crazy, media remove alone removed 42GB. The console output said 37GB, but I checked my dataset and it went from 50GB to 8GB
Thanks a lot!
Show threadSanderFeb 28
@marco cool, nice to see you could indeed free up space.