Moodle 2.0 File Storage

One of the major re-engineering efforts undertaken for Moodle 2.0 was a complete redesign of the Moodle file storage system. The release of Moodle 2.0 has delivered a database centric file handling model, replacing the OS file system centric model.

Moodle <= 1.9 file storage

Moodle stores files in a number of circumstances: student assignment submissions and teacher course resources being two examples. How would you store them? In Moodle 1.9 a directory structure was used to identify the files, using a combination of the course, module and user ID numbers. So, in your Moodle data directory you would see a number of folders corresponding to Moodle course ID numbers, in which you would find the course files (if any) and a ‘moddata’ folder.

The ‘moddata’ folder contains sub folders corresponding to the Moodle module types, with further sub folders corresponding to the module instance ID numbers, and yet further sub folders corresponding to the user ID numbers which contain the user uploads to that instance of that module.

Here is an example file and folder structure for a student with ID number 123 uploading an assignment submission to an assignment instance with ID number 456, in a course with ID number 789:

 {MOODLEDATA}/789/moddata/assignment/5/123/my_midterm_assignment.doc

Corresponding database records are stored on a per module basis (if stored at all). Assignment submission details like submission date are stored in the database table ‘assignment_submissions’. Make sense? Perhaps, but it’s not ideal. This configuration makes backup and restoration more difficult for a start.  This method also means that the same file could be stored multiple times in different parts of the Moodle data – something which is likely to happen if you copy and re-use courses from year to year.

Moodle 2.0 file storage

Moodle 2 centralises the storage of files within a {MOODLEDATA}/filedir directory and a single ‘files’ database table. A new files API has been created to provide higher level of control. All files are referenced by a record in a files table and accessed via the files API.

In the background, files uploaded into Moodle 2.0 are given an SHA1 hash value, and this hash value is used as the basis for storing the file in MOODLEDATA/filedir/HASH_PATH/HASH_FILE.

In the example {MOODLEDATA}/filedir/5c/eb/b0/5cebb03d702827bb9e25b38b06910fa5 the hashed value for the file (5cebb03d702827bb9e25b38b06910fa5) is stored in a hash path created from the first three pairs of characters in the hash value: [5c][eb][b0]3d702827bb9e25b38b06910fa5. This method of storage means the same file uploaded multiple times will only be stored once, because the hash value will be the same. Using the hash value as a basis of creating the file storage directories creates a natural distribution of the files across multiple folders. This helps reduce the possibility that any one directory will contain an excessive number of files.

There are other benefits to this file storage method; backup & restores are simplified, and all file access is through the file storage API giving better control. Additionally, the original file names are not recorded on disk which allows for full Unicode file name support, regardless of the Unicode support in the underlying operating system file system.

One possible downside is that you will no longer be able to look at the Moodle data directory and immediately understand and modify the files with ease, but should anyone be doing this anyway?