Try this (as root):
mkdir -p /etc/systemd/system/uwsgi.service.d
cat > /etc/systemd/system/uwsgi.service.d/old-exec-start.conf <<'EOF'
ExecStart=/usr/sbin/uwsgi --ini /etc/uwsgi.ini
and then (still as root):
systemctl restart uwsgi.service
Why did this happen? Shortly before the release of CentOS 7.8 (so it was not caused by that update itself - it’s just the time where people are most likely to actually restart uWSGI or even reboot the server), an updated uWSGI package has been released which changed the start command for uWSGI:
-ExecStart=/usr/sbin/uwsgi --ini /etc/uwsgi.ini
+ExecStart=/usr/sbin/uwsgi --ini /etc/uwsgi.ini --pidfile /run/uwsgi/uwsgi.pid --stats /run/uwsgi/stats.sock
When running it was the default user this would not be a problem, but since we are running uWSGI as the indico user, it now fails to create the stats socket file in
/run/wsgi. By adding the custom config file in
uwsgi.service.d this change is basically reverted, and everything works again.