TLD
The domain clarin.eu is registered by the MPI and will be transferred to the CLARIN-ERIC legal entity when it is in place.
Right now the following subdomains are in place:
clarin-eu subdomains:
- www.clarin.eu > clarin.mpi.nl (= 192.87.79.165)
- catalog.clarin.eu > 192.87.79.171
- svn.clarin.eu > 192.87.79.165
- trac.clarin.eu > 192.87.79.165
national subdomains:
- de.clarin.eu > 141.58.160.11 (contact: Jens Stegmann, jens.stegmann@googlemail.com)
Migration of the site to new server
Move the current installation to a new virtual machine, featuring:
Installation details
URL: http://www.clarin.eu/internal
(www.clarin.eu, the external site, runs on a set of php pages on the samen web server)
webserver: clarin.mpi.nl - virtual ESX host at the MPI, running ubuntu 12.04 LTS (supported until 2017)
software running: LAMP, drupal 6
apache root: /var/www/drupal
The user database is also exported to the "home for the homeless" IdP that is running at the MPI. The latter is used e.g. for the Component Registry and thus forms an important part of the infrastructure.
/var/www/drupal/external/language.php provides the entry pages for an ISO-639-3 code, used by the Google Earth file and to be used by the VLO, eg: http://www.clarin.eu/external/language.php?code=nld
Patch applied for www.clarin.eu/export_resources : http://drupal.org/node/366120#comment-3558700
Known issues
- ticket:31 , related: currently for the hierarchical structure of the site there is a mix between the standard book pages (http://www.clarin.eu/admin/content/book) structure and the non-standard site outline (http://www.clarin.eu/content/hierarchy). Only one mechanism should be used (probably the book pages as that is a part of the drupal core and thus probably better supported in the future). This is a non-trivial task but very important nevertheless as the current situation obfuscates the structure behind all information on the site.
- ticket:39
- ticket:33
- there is no CCK node reference field in event that links to the presentations that are related
- The access control mechanism via organic groups is too complicated. A few roles (anonymous, clarin member, clarin EB) should be sufficient. Again going from OG to roles is not really trivial.
Other scripts on the same server
- /home/dietuyt/bin/drupal-cron.sh: shell script that calls the scripts that need to be run every hour
- /home/dietuyt/bin/org2clarin.py: script that generates HTML out of http://www.clarin.eu/export_organisations and updates /var/www/drupal/exernal/data/clarin-organisations.txt which is used a source for the member pages at http://www.clarin.eu/external/index.php?page=about-clarin&sub=3&sub2=0 with the most recent information from the organisation database at the internal site. Should be run every hour by cron but for some strange reason this does not happen. Should be run (as a workaround) after each addition of a new organisation (http://www.clarin.eu/node/add/organisation).
- /home/dietuyt/bin/update-cmd.sh: checks out the latest version of the metadata directory in SVN (source:metadata/trunk/toolkit) to www.clarin.eu/cmd - this path is used by the ComponentRegistry which relies on it for eg comp2schema.xsl
