>... or just doing the full-backup with dump, too and splitting >the archive "manually" (by an external script) to 3 tapes. But >how will amanda know "never do full-backup, not even an initial" >on this partition? How will amanda "know" about the existing >full-backup and will start(!) here with incrementals and will >never request a full-backup for this partition?
The online book chapter (see the bottom of the www.amanda.org page) covers this, although I don't know that I tested it very much. Here's what it says: To run full dumps by hand outside of AMANDA (perhaps they are too large for the normal tape capacity, or need special processing), create a new dumptype and set strategy to incronly: define full-too-big { comp-user-tar strategy incronly } Tell AMANDA when a full dump of the area has been done with the force suboption of amadmin. Take care to do full dumps often enough that the tape cycle does not wrap around and overwrite the last good non-full backups. To never do full dumps (such as an area easily regenerated from vendor media), create a new dumptype and set strategy to nofull: define man-pages { comp-user-tar strategy nofull } Only level 1 backups of such areas are done, so wrapping around the tape cycle is not a problem. The different between the two is that incronly will go ahead and bump to higher incremental levels (2, 3, and so on) while nofull will stay at level 1 forever. Note the following warning about incronly in amanda(8): Unfortunately, it appears that Amanda will perform full backups with this configuration, which is probably a bug I don't remember if this has been fixed. And it may not be a problem in your case since the level 0's are very large. As to the initial full dump Amanda will insist on, I'm not sure if using "incronly" and "force" will get around that or not. If it doesn't, a nasty trick would be to: * edit down the disklist to just the file system in question * change your tapetype to a very large size * change your tapedev to /dev/null (or null:, depending on the version of Amanda) * run amdump and let Amanda do its thing, throwing away the resulting image * put everything back the way it was At this point, Amanda should be happy, thinking it has done the initial full dump, and one of the two other techniques above should work. Obviously, this should all be cleaned up some to make it easier to do. It comes up often enough that it should be better supported. >Raphael Becker John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]