rsk@mace.cc.purdue.edu (Rich Kulawiec) (11/16/88)
I just wanted to take the opportunity to plug the entombing system developed locally, and now in general use here. For further info, it's probably best to contact Matthew Bradburn at mjb@j.cc.purdue.edu, since he's the author. Enclosed below is the abstract from the paper describing the system. "rm * .o" vs. the Novice User Matthew J. Bradburn Purdue University Computing Center West Lafayette, IN 47907 ABSTRACT This paper addresses the need for a mechanism to recover accidentally deleted files. We consider various procedures for saving ("entombing") the these files and the relative merits of each. Many of our users occasionally have reason to require the restoration of some file, or group of files, which has been accidentally deleted. Currently, each user must create his own file backup procedure which may not be efficient, effective, or secure, and which may cause some inconvenience by consuming a large portion of the user's allotted disk space, or they must burden the Computing Center staff with requests to retrieve files from backup tape, a procedure which may involve the loss of many hours' work and which may not result in the recovery of the file. While we recognize that the most comprehensive method of solving this problem is to do so at the kernel level, we feel that practical considera- tions (including portability) make other solutions more attractive. Our system of file entombing can be added to any program without source code changes by recompiling it with our entombing library. This causes our subroutines to be sub- stituted for system calls which commonly destroy files.