This report takes a long time because it basically traverses all the repository and calculates the size of the nodes and properties. Depending on the amount of content and the repository architecture (tarmk, s3 datastore, mongo, ....) it can take a significant amount of time. Anyway, I would never start at "/". But it will never report an accurate number of consumed megabytes and gigabytes on disk, mostly because it does not understand the underlaying architecture. For example, if you have a datastore, it will not take the deduplication effect of the datastore into account. Also it does not account for indexes.
To get the relevant numbers of disk consumption (as a total value), you should have a look into the filesystem. It cannot give you details in which area of the repository the largest pieces are stored, but it gives only the overall number. For a deeper level of granularity you have to use the disk usage report.
Jörg
