CLAYTON@XRT.UPENN.EDU ("Clayton, Paul D.") (07/25/87)
Information From TSO Financial - The Saga Continues... Chapter 14 - July 25, 1987 Todd Warnock has asked the following question about BACKUP and implications with volume sets. While it is NOT supported (that I know of), has anyone had experience with using a PHYSICAL restore (BACKUP/PHYSICAL) to restore a member of a volume set without restoring the whole thing ? Assume that the volume set is logging errors at an extremely high rate (but is still functioning). The volume set is dismounted, the bad disk (disk two of a three volume set) is PHYSICALly backed up to tape, a new HDA is installed, and the PHYSICAL backup restored. I would not recommend that PHYSICAL backups be done to volume sets for the specific volumes in a set. Considering that a RA type disk on an HSC/UDA is SUPPOSED to present a 'perfect and error free' picture to the host, it would probably work. On NON RA devices you are cruising for a bruising. Another approach I think might be better is to do a /IMAGE with the /VOLUME=n, where n is the volume number in a set. By doing this the disk is also being compressed and reorganized to provide for more contigious space for future expansion(s). The payback I think is worth it. Anytime I have to copy 800K to 1M blocks of information, I do what can be done to put the volume in the 'best' shape. I do not use any of the third party packages to 'place' files. The only special case is for initializing a disk and using the /INDEX=(begin,middle,end,block:n) qualifier to place the INDEXF.SYS file on the disk. This would be coupled with the /NOINITIALIZE qualifier on the BACKUP command stream. Hope this helps. :-) Paul D. Clayton - Manager Of Systems TSO Financial - Horsham, Pa. USA Address - CLAYTON%XRT@CIS.UPENN.EDU