Because the resources required for backups are significant, we strongly encourage users to exclude any data that does not require backup. Some examples of such data are:
Any data that can be retrieved from internet repositories
Any data that can be easily and reasonably quickly regenerated via computation
Any data that is temporary or intermediate data generated by computation
Any directory (and it's nested files and subdirectories) can be excluded from backup by simply including the string "nobackup" (all lowercase) anywhere in the directory's name. Additionally, any individual file with the string "nobackup" (all lowercase) anywhere in it's name will be excluded.
For example, the following directories (and all their contents and subdirectories) will not be backed up:
/proj/myproject/dataset1.nobackup/
/proj/myproject/dataset2/nobackup_subdata1/
/home/me/matlabresults/resultsnobackup/
/home/me/nobackup/
/home/me/data_crunching/nobackup.temp