What is a Wikimedia Backup?
A wikimedia backup contains all of a wiki’s data in an XML file. This includes pages, revisions, discussion pages and files (but not user data).
This is useful for emergencies such as a deliberate internet blackout by activists or unexpected problems such as power outages. It can also help prevent censorship.
Database backup
A database backup is a file containing all the data in the wiki’s MySQL/MariaDB database. It includes pages and their revisions, discussions, user accounts, image galleries (including deleted images), wikitext markup, metadata, and search index. It does not include user passwords, email addresses, or any other personal information.
A wikispace owner can create a database backup using the space tools menu, or by visiting Special:DataDump on their own wiki. The wiki administrator can also request a new database dump by clicking the “Send request” button, which will update the date on the download options.
It is recommended that users back up their wikis regularly, even if they never need to restore them. A good time to make a backup is after installing an extension such as DeletedImagesRestore or MediawikiDumpToStaticHTML, which will automatically generate an HTML version of your wiki for offline reading. This can be especially useful in case of deliberate blackouts such as the January 2012 protest of SOPA and PIPA.
File backup
Unlike pages, media files are stored as regular files on the file system rather than as blobs in the database. As such, a complete backup also includes the media directory (usually named wiki/images/) as well as derived thumbnails in the thumbs/ folder. These are stored separately from the XML database dumps and can be archived using simple *nix commands. The size of the media archive is largely dependent on the size of the wiki and its image files; for example, a full English Wikipedia dump requires about 7.5 gigabytes of space and can take several hours to download over a typical connection.
ITS captures a file backup of the entire wiki environment and individual spaces once every 21 days. If you need to restore an individual space after that, you must make your own backup. This is important because any work done on your space after the backup was taken will be irretrievable. The backup contains the XML file for all current pages and all page history. The file may be large for communities with many pages or revisions.
Image backup
The MediaWiki software engineers decided that images should be stored as regular files on the file system, rather than BLOBs in the database. So backing up the database does not back up your wiki’s images, which should be backed up separately.
If your wiki is large, you might want to consider backing up all the data and files in an unstructured repository (such as tapes, DVD-Rs or external hard drives). This method requires a significant amount of manual intervention.
It can take a while to restore an image backup, depending on the size of your wiki and how much you have backed up. It is best to practice restoring your backups, to become familiar with the process and prevent future anguish. There is a cross-platform program called XOWA that can be used to view and render wiki dumps.
Restore
A wiki contains many different types of data. It is important to backup these files regularly, so that you can restore your data if needed. You can use the wikimedia backup tool to make these backups. The wikimedia backup tool can also be used to move your wiki from one server to another.
You should also backup your file system components, e.g., images, extensions, and custom skins. This can be a time-consuming task, depending on the size of your wiki.
To automate the backup process, you can use a cron job. Cron is a time-based job scheduler in Unix-like computer operating systems. The following example shows a script that runs every day at 5:10 AM. It checks a directory for backups and restores them if necessary. The backup files are then stored in a special folder. The script also creates a symlink from each backup to the new version of the wiki. This ensures that the latest changes are backed up, even when the original database is updated.