These steps should be done roughly every day for now, until automated fully.
All servers run Ubuntu and rely primarily on native-distro packages for all services. Server maintainance is as expected for non-commercial Ubuntu.
Log into the server and run these commands:
sudo apt-get update
sudo apt-get upgrade -y
Optionally reboot if necessary, or every month at the longest.
Currently the mail for root and aws-backup is not delivered through SES. Check these manually every few days using the mail command.
In the future I may forward this elsewhere using
procmail.
TLS certificates are issued through LetEncrypt. A cron job attempts to rotate this certificate every day, if it is near expiration.
$ cat /etc/cron.daily/letsencrypt-renew
#!/usr/bin/env bash
certbot renew
systemctl reload nginx
The certificate was last set to expire on February 23rd, 2023. Check that it does indeed renew before then.
There appears to be a slow memory leak in the Sidekiq container causing it to consume more than 800MB of memory. Recycle the container every day or two until this can be automated.
$ sudo su - mastodon
$ cd live/
$ docker-compose stop sidekiq
Stopping live_sidekiq_1 ... done
$ docker-compose start sidekiq
Starting sidekiq ... done
Cleaning-up cached federated content is done from the mastodon user's crontab:
10 19 * * * docker exec live_web_1 bin/tootctl media remove --days 7
5 19 * * * docker exec live_web_1 bin/tootctl preview_cards remove --days 7