How do I create a group-based disk usage script?

Dave's Answer: First off, you say that the users are sharing a particular directory, so I'm going to assume you mean that they all have the same home directory on the system, but have discrete entries in the password file (that is, different user IDs). This would be shown if you did an ls -l and saw files owned by each of the users in the same home directory.

That being the case, I'll first suggest that you put them into different home directories. You could even have those directories under the shared directory, as in /home/group/A and /home/group/B, etc. That'd make things easy to track. Failing that, it's still quite possible to ascertain disk usage on a per userID basis by using the find command.

For example, consider using find / -user A -type f -print, which would generate a list of all files owned by user A across the entire file system, or, better, using the slightly more complex find / -user A -type f -exec ls -ld \{\} \; which generates output similar to:

-rw-------  1 taylor  staff   21508  1 Nov 23:20 ./.DS_Store
-rw-r--r--  1 taylor  staff       0  8 Apr  2004 ./.localized
-rwxr-xr-x  1 taylor  staff  950272 29 May  2003 ./021022_0801.bin
-rwxr-xr-x  1 taylor  staff      69 17 Sep  2003 ./021022_0801.cue

With that information, you should be able to create a brief awk script that sums up all the numbers in column #5, divides by 1024 and presents the final disk space used by each user in the group in Kb (or divide by 1024*1024 and get Mb). To go from that to calculating what % of the total alloted space for the group each person is using should be straightforward.

Hope that helps you!