SBAILEY@dominoes.UUCP.UUCP (05/14/86)
First, a quick comment about the HSC-50 backup question that's been floating around the network... In article <8605121408.AA01209@uwmcsd1.UUCP> Sandy Berger writes: >In article <8605102347.AA29393@ucbvax.Berkeley.EDU> >jbredese@CARLETON.CSNET (John Bredesen) writes: > >>1a) If so, is the tape in backup format so it can be used by the host later >> on? > >The format is unique to the HSC and cannot be used by the host. Files can not >be restored on a selective basis. It's all or nothing and the HSC must be used >for the restore. This is not the case. The tape is written using the same format that VMS BACKUP uses when you do a /PHYSICAL backup. In particular, if you have a save set that was created by the HSC-50 you can: $ MOUNT/FOREIGN/OVER=ACC $1$MUA0: $ BACKUP/PHYSICAL $1$MUA0:/SAVE $1$DUA51: or whatever. If you like, you can BACKUP/LIST and see that some of the fields in the header are scrambled, but the important stuff (name and block count) is still there. I tried doing a restore like above and it worked for the 1st half-reel at least. (I got tired of waiting and didn't want to do the whole 3.5 reels.) I don't know what you'd have to do if you wanted to restore a VMS backup using the HSC-50; probably something similar to: $ BACKUP/PHYSICAL DISK: TAPE:SAVESET/REWIND/NOCRC ! Off-the-wall guess Now that the record is straight on that, I have a question about stand-alone backups: "WHY?" We routinely do the usual VMS BACKUP/IMAGE, etc. on our system disks as well as our user disks. I do not worry about the few files (ACCOUNTNG.DAT, etc.) which are open for write on the disk since I can live without 100% of my accounting or error log or whatever data. In fact, I have been making private system disks for new nodes in our cluster by BACKUP/IMAGE from an active system disk to a spare disk, and then changing the SYSn directory name to whatever I want, before proceeding to hacking swap/page files, startup procedures, etc. Is there something inherently better about standalone backups (besides the pain) that I should know about? Unless we get mauled by an earthquake or similar disaster, one of our nodes is always up and available for doing saves/restores of other nodes, and even if we had to rebuild from scratch, my impression is that I could use our normal image saves and restore them using standalone backup. If I'm wrong, I'd really love to find out about it while I can still recover gracefully. Thanks in advance for the explanations. Scott Bailey UUCP: {scgvaxd,cit-vax}!engvax!dominoes!sbailey VAX System Manager Xerox: sbailey:ES GSD/WCO:Xerox Xerox Corp. RE/GSD/WCO Phone: (213) 536-5441