D4Science Infrastructure

Before MongoDB (the actual storage used behind the workspace), files were stored into jackrabbit itself (through the DataStore facility). Some GBs of no longer used files need to be freed up. In order to do so:

• a script must be executed to replace the payload (if any) of file nodes with an empty payload (a background job that can be executed with workspace up and running);
• the DataStore Garbage Collector must be executed (it needs the workspace down, so we need to schedule its execution properly and extimate how long it takes to finish)

We tested once this operations onto a snapshot of the current content of jackrabbit in production and we were able to free up 25GB of files. However, we need to better extimate the time needed for the whole task to finish.

It is important because will let us migrate the datastore into postgres (no need of shared file system if we plan to have replicated jcr instances) and speedup the migration phase