Blueskying here: Backing up amanda's data has always been a problem.
Gene's solution of running, after amdump, a separate, backup program and adding it to the end of the last amanda tape could work, but a method within amanda itself would be the better solution. Gene's approach also has a big advantage of the data not changing while it was being backed up. I wonder, do the new changes and APIs would allow us to revisit this "problem". I'm thinking of something like a new dumptype "run last" or "run after". The DLE it is applied to could list the server's root directory as the starting point and have several "include" directives to specify the several amanda directories to backup (even if on different filesystems). The above approach would save the amanda data as the last data on the media. Two alternative approaches would put the data at the head of the next dump. One would be to run an second amdump after the primary one completed that dumped only the one DLE and left it on the holding disk to be auto-flushed later. The other approach would be a "run before" dumptype. Any thoughts on how desireable you feel be a separate copy of amanda data would be and other approaches? Jon -- Jon H. LaBadie [email protected] JG Computing 12027 Creekbend Drive (703) 787-0884 Reston, VA 20194 (703) 787-0922 (fax)
