johnston@lbl-csam.arpa (Bill Johnston [csr]) (01/05/90)
Unix-Wizards; I am looking for file archive software, or pieces that I might start from to make an archive system. What I want to do is more or less the following: Have a user cmd (archive) that marks a file as a candidate for archiving. ``mark'' probably means make an entry in an archive DB file somewhere that do_archive reads. Have a (probably cron run) command (do_archive) that checks the archive DB for archivable files, writes those files to tape, and keeps track of where they were put (updates the archive DB). ``keeping track'' of where they were put means what tape, and where on the tape. This is somewhat tricky as it means a labled tape format (though the label could be a text string in the first file of the tape). It would be nice if the main part of the tape format were standard (e.g. tar). Have a recover cmd (that uses the archive DB information written by do_archive) to retrieve files without the user having to know what tape, and what file on the tape has the (tar) archive containing the required image file. The ``real'' file name (as kept in the archive DB) might be something like a full path name, but an alias mechanism would also be nice. At the least I have seen potentially useful ``pieces'' pass by on the source archives over the years: seems that there was a version of dump that records files deleted since last dump (if I remember correctly), and I seem to remember an ``incremental'' tar of some sort. Pointers to these would be appreciated. On a related topic, we are archiving images, and I am seriously considering DAT instead of Exabyte type video format tape. The reason for this is that I understand that access is much faster for DAT, and we may be shuffling files off of and on to tape frequently. Anyone have any experience with the DAT drives, software, etc? Any problem with skipping to the end of previously written tape and adding files? (We will probably have several of these drives on Sun SpacrStations.) I do not normally read this list, so send comments directly to me if you can: Internet: wejohnston@lbl.gov uucp probably is: .....ucbvax!csam.lbl.gov!johnston Thanks, Bill Johnston Human Genome Center Lawrence Berkeley Laboratory