Notes about CDF Raw Data Tape Logging and SAM Uploading

Tape recycling is a very case-dependent process. It can be roughly summarized to one or more iterations of the following steps:

1. identify the files to delete, either their names (dataset names in CDF, or from database query in D0) or patterns in their names;
2. get the Enstore complete file listing for CDF to a local disk using either "wget" or "curl" -- These listings are formatted text files about 300 MB in size; they contain all the active -- not yet deleted from pnfs -- files;
3. parse the comlete file listing to find the (inclusive) list of tapes (the 3rd column) which contain at least one active file satisfying the deletion condition;
4. set up the proper version of encp, and do "enstore volume --list ***" to get the list of all files on each candidate tape -- including files already deleted from pnfs but still on the tape;
5. check to be sure that all the files on any particular tape, active or not, satisfy the deletion condition or should have been deleted before;
6. login as a proper user (usually "sam") to a machine where pnfs is mounted with write permission, delete the pnfs entries, and mark the files "bad" in the SAM database;
7. re-check the tapes with "enstore volume --list ***" to be sure that there are no active files on the tapes to be recycled; send the list to enstore-admin.

Please note that when enstore-admin gets to do the actual recycling, they may reject a few tapes if new files have made to those tapes after the last check. This is normal.