This project holds the website for the An Crúbadán project, built
on the clld framework.
This documentation was written by Nick Lewchenko. Questions about it should be directed to him.
This server depends on the following utilities:
-
git
-
virtualenv (Ubuntu package:
python-virtualenv) -
python2.7
-
a webserver, probably
nginx
The clld server process should be run as an unprivileged user, so create one for that purpose if you don't have one.
The main data directory, /data/crubadan-clld must be in place
and read/writable for the clld server user from the last step.
In the main data directory, you must also create the files/,
metadata/, and olac/olac.xml files and directories and
fill them (this should be all taken care of by the scripts that run
the core An Crúbadán services).
It doesn't matter where you put these, but do note that they have to remain in place while the service is running. A sensible place would be the new user's home directory.
Make sure that you are running this (and all commands given from here-on in the document) as the user who will run the server:
~ $ git clone http://github.com/kscanne/crubadan_clld
~ $ cd crubadan_clld
You should now be in the sources' root directory.
Copy ./crubadan_clld.conf to /etc/nginx/conf.d/ (This step
will require root privileges, of course).
This configuration will probably need some adjustments to work on your
system. I'm no nginx configuration master... If there are any
general rookie mistakes, please pull-request.
This server has many python dependencies not listed above, but they are meant to be installed into a local sandbox with one command:
~/crubadan_clld $ ./sandbox-init
This sets up a sandboxed python installation in .venv. Commands
given to ./service-control will use this behind the scenes.
Before starting the server, build its database:
~/crubadan_clld $ ./service-control build
This will write into /data/crubadan-clld, which again should be
writable with your unprivileged server-user.
This same command should be run any time there are updates to the data available in the project. Doing so does not require a restart of the server processes.
Make sure that nginx is updated and running, and then start the clld server processes:
~/crubadan_clld $ ./service-control start
If this goes wrong, or seems to have, check if the pid files were
created and whether any pserve processes are running (there
should be two of each):
~/crubadan_clld $ ls | grep pid
~/crubadan_clld $ ps -e | grep pserve
The server should now be running! Navigate a web browser to the hostname you've set up and see if it's there.
The clld server processes can be stopped with:
~/crubadan_clld $ ./service-control stop
The same troubleshooting steps given for start can be used here
if it seems to fail.
If upstream changes are made, just stop the server, use a git-pull to update the sources in place, and rebuild the sandbox and database before starting again.
~/crubadan_clld $ ./service-control stop
~/crubadan_clld $ git pull
~/crubadan_clld $ rm -r .venv
~/crubadan_clld $ ./sandbox-init
~/crubadan_clld $ ./service-control build
~/crubadan_clld $ ./service-control start