Snapshots: Restores on the NetApp are done from snapshots. Our default snapshot policy is to do snapshots every 4 hours (called an "hourly" snapshot). Hourly snapshots are kept for 2 days. The snapshot done at midnight is called a "nightly" snapshot. Nightly snapshots are kept for 2 weeks. And the snapshot done at midnight Sunday night (Monday morning) is called a "weekly" snapshot. Weekly snapshots are kept for 2 months. So under our default policy, there are about 30 snapshots of the filesystem from various times over a period of 2 months should a file restore be needed. If you like, you will be able to specify a different snapshot schedule for your project space.
Exports: Project space can be exported (made available for mounting) to LCSR administered machines, self-administered project lab machines, or both. NB: if you export your filesystem to both LCSR machines and lab machines, you may end up with ownership/access problems unless you coordinate your UID/GID spaces with LCSR.
So a simple request for a project qtree might look like this:
To: help@cs Subject: Project space request Please create a project qtree to be mounted as /koko/example on the faculty and research clusters. It should initially be 20G in size and writable by the cs-bozo group with the default snapshot schedule.while a request for more project space might look like this:
To: help@cs Subject: Project space request Please add 20G to the space allocated to /koko/example.Requests for project space allocation are subject to space available.
Other options for space on the DCS faculty and research clusters are described here.
Addendum added 3/18/16: The numbers in this policy may be adjusted for our new NetApp installed in 2015. For details on the information I need to create a new qtree on the new NetApp, see my updated list of specifications.