home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.vms
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!darwin.sura.net!Sirius.dfn.de!gwdu03.gwdg.de!gwdgv1.gwdg.de!MOELLER
- From: moeller@gwdgv1.gwdg.de
- Subject: Re: re^2: extending INDEXF.SYS (new idea!)
- Message-ID: <AA48Q3E@gwdu03.gwdg.de>
- Sender: news@gwdu03.gwdg.de (USENET News System)
- Nntp-Posting-Host: gwdgv1.gwdg.de
- Reply-To: moeller@gwdgv1.gwdg.de
- Organization: GWDG Goettingen, F.R.Germany
- References: <9211170506.AA25540@ucbvax.Berkeley.EDU>
- Date: Sat, 21 Nov 1992 00:45:32 GMT
- Lines: 46
-
- In article <9211170506.AA25540@ucbvax.Berkeley.EDU>, I wrote:
- >Some time ago, there was some discussion re: pre-extending INDEXF.SYS,
- >and the consensus was, that one has to be very careful,
- >so that INDEXF.SYS won't become a multi-header file by any chance.
- >
- >It occurred to me that there is a VMS utility for extending files,
- >which *has* that important check built-in:
- >
- > $ MCR SYSGEN CREATE disk:[000000]INDEXF.SYS/size=<newsize>
- >[...]
-
- Sorry, it wasn't a good idea - please forget it!
-
- Basically SYSGEN CREATE does the right thing by extending a file with
- FIB$M_NOEXTHDR, which avoids getting a multi-header file.
-
- But there several problems:
-
- A minor one was noted by D.Jones, in that any extend operation
- probably won't update the alternate index file header
- (which allows recovery from a damaged primary header);
- ANALYZE/DISK/REPAIR should fix that - often that's a good idea anyway.
-
- Unfortunately, SYSGEN does two more things:
-
- (a) the equivalent of SET FILE/NOBACKUP (noted by simons/G=Colin/...)
- I doubt that this would be a problem, because BACKUP saves INDEXF.SYS
- "file header only" anyway (look at your BACKUP Journal!).
-
- (b) the equivalent of SET FILE/END_OF_FILE, i.e. it explicitly sets
- the logical end-of-file mark to the highest block allocated
- This isn't desirable anyway (makes INDEXF scans slow);
- unless the disk is being used for the first time,
- there's also a good chance for some of the blocks
- formerly beyond e-o-f in fact to look like valid file headers
- (from a previous copy of INDEXF.SYS, which usually occupies
- the same disk blocks after a re-INITialization).
- Run ANALYZE/DISK/REPAIR, and you suddenly have lots of
- "multiply allocated blocks" (due to re-appearing "lost" files)!
-
- Just yesterday I received VMS 5.5-2; if BACKUP/NOINIT now finally does
- the Right Thing, exteding INDEXF.SYS may become a topic of the past.
-
- Wolfgang J. Moeller, GWDG, D-3400 Goettingen, F.R.Germany | Disclaimer ...
- PSI%(0262)45050352008::MOELLER Phone: +49 551 201516 | No claim intended!
- Internet: moeller@gwdgv1.dnet.gwdg.de | This space intentionally left blank.
-