Backups are handled elsewhere now

Script ran fine on the console, failed in cron job.  This is because
the docker-compose exec command was attempting (by default) to allocate
 a TTY, and failing.  This was also not being detected because the
default bash behaviour is to return the status of the last command
in a pipe (in this case, gzip). Enable pipefail option to prevent this.

And on runko.social.coop.  Used as a stopgap backup solution there.

A stopgap until we get a proper backup

Copied from the cryptpad version.