XML Site Backups

As of version 1.1, Segue supports the exporting of individual or all sites to XML files via a collection of command-line PHP scripts located in the segue/export/ directory. The export scripts can be run manually, or they can be schedualed by cron or another schedualing utility. These XML dumps can be imported back into segue using the import_site.php script.

Problem:
A problem arrises when the restoration of the database from backup after a server failure results in a corrupt or missing site. This corruption may go unnoticed for some time while other sites are edited/added to the database. Restoring this site from a database dump (and ensuring valid keys) would be a time-consuming job.

Solution:
It is recommended that database backups be done in conjunction with these XML backups. Exporting sites to XML files which preserve the relationships between elements without the use of keys allows restoration of individual sites with new, valid, key linkages and much automation.

Individual sites can be exported as an XML file and subdirectory of media files using the export_site.php script. This script takes the name of the site that is desired to be exported as well as an existing output directory. Ex:

Be sure to run this script as a user who has permission to access the Segue  media directory or media files will not be copied. For more help, run:

More often than one site, you will probably want to export all Segue sites on a regular basis. To accomplish this, use the export_all_sites.php script. This script takes an existing output directory as a parameter. Ex:

Be sure to run this script as a user who has permission to access the Segue media directory or media files will not be copied. If you would rather have a compressed archive put in the output directory, use the "-z", compress, option.

The "-z" option will export sites to /tmp/segue_tmp/, then create a allsites.tar.gz file in <output directory>. This option requires the ability to execute a command of the form "tar -czf /backups/segue/allsites.tar.gz /tmp/segue_tmp/". For more help, run:

Individual sites can be imported into Segue from an XML file and directory of media files. This can be used to restore a corrupted or deleted site from an XML backup. There are several expectations that the import script has of the XML file:

• All users refered to in the XML file exist in the destination Segue instance.
• All [media] files refered to in the XML exist in the specified source directory.
• The site does not exist in the destination Segue instance.

The goal for the initial implementation of these scripts was to support backups and restoration of sites upon user request. The storage of user, group and class information necessary for accurate user restoration is beyond the scope of this goal. If users don’t/can’t exist in the destination Segue, remove references to them in the XML file if you still wish to import the site.

If media files do not exist in the source directory, the script with fail with notices. Either put the files in the directory or remove references to them in the XML file. The script will not guess which is appropriate.

This script will not overwrite existing sites. Please delete the existing site first before importing.

Assuming the above conditions are met, import the site with a command such as the following:

For more help, run: