I’m trying to troubleshoot an installation of LiquidFeedback (LF) on a VPS for, hmm, the Israeli Pirate Party? Anyway, aiming to DTSTTCPW, but the process is (quite) a bit ugly, so I needed a place to document stuff…
- We’re on a VPS provided by GPLHost, running Debian.
- LF provides installation instructions. Shuki, the sysadmin, did everything already.
- But, the instructions are broken/incomplete:
Start sending of event notifications: (TODO: daemonize) cd /opt/liquid_feedback_frontend/ echo "Event:send_notifications_loop()" | ../webmcp/bin/webmcp_shell myconfig
neglects to say webmcp_shell must be run as user www-data, not root, or else:
lua: /opt/liquid_feedback_frontend/config/init.lua:46: FATAL: Ident authentication failed for user "root"
- And the “TODO” note? Yes, this script never returns. Daemonize it already, please, someone?
- Googling… meh.
- Note the Canadians use Apache/CGI and seem to have neither of these problems. The Germans do use Lighty, but don’t mention this WebMCP “Event:send_notifications_loop()” daemon? Foo?
- And after that, they say, restart the webserver:
Smells like trouble. But, nu.
- Should I write to the (unsearchable!) mailing list to complain? The issue tracking system it seems only deals with code.
Starting web server: lighttpd2012-07-12 00:11:27: (plugin.c.131) Cannot load plugin mod_redirect more than once, please fix your config (we may not accept such configs in future releases .
[sic] Fixed by commenting it out in /etc/lighttpd/conf-enabled/60-liquidfeedback.conf.
- Lighty wasn’t configured for virtual hosting, so, for now, site is reachable at http://22.214.171.124/lf/. All of LF’s URLs seem prefixed with /lf/, but still, that’s not how things ought to be done.