According the changelog, 2019-11-27 entry says "07:13 dereckson: Disk drives look corrupted. Attach a new 1 To drive to the machine. Install a clean CentOS 7 on this drive to recover data."
Per T1592, we're now on CentOS 8, and as Dwellers is the Docker development and staging server, we're going to use CentOS 8 stream.
Server has been reprovisionned, there is still to apply highstate, and review pending Dwellers Maniphest tasks.
Errors to fix
- nginx - install nginx instead of only configure it?
- nginx - provide base nginx.conf
- nginx - provide roles/paas-docker/nginx/files/vhosts/notifications.conf
- nginx - provide roles/paas-docker/nginx/files/vhosts/bugzilla.conf
- nginx - provide 502 on fallback while we're done?
- nginx - use wildcard certificates for fallback?
- nginx - modernize ciphers per [https://ssl-config.mozilla.org/#server=nginx&version=1.14.1&config=intermediate&openssl=1.1.1c&ocsp=false&guideline=5.4 this mozssl configuration]
- TLS - crontab with certbot renew --post-hook "nginx -t && nginx -s reload" (but see Sandlayth code, it's more complicated)