[Info-vax] Problems detected with analyze/disk

[Peter Weaver]

> > > The problem in this case seems to get rid of files with "unprintable"
> > > characters
> > > in their names, with just access to the file-id. With a full file spec
> > > I could have
> > > used the recommended procedure to copy all multiple allocated files
> > > and delete
> > > all but one, then repair the disk. Just don't know how to do it fith
> > > only file-id's.
> > 
> > Since you have V8.3 and ODS5 you can do the following;
> > 
> > $ DELETE/CONFIRM/LOG SYS$SYSDEVICE:[0000000...]~[708146,53,0]
> > $ DELETE/CONFIRM/LOG SYS$SYSDEVICE:[0000000...]~[156375,23,0]
> >
> > If the files are not on SYS$SYSDEVICE then substitute your disk name,
the
> > "~" character is needed and you do not need any "*" characters in the
> > command. The confirm message will show you the device, directory,
filename
> > followed by the FID in square brackets then the extension and version
> > number.
> > 
>
> Rather than deleting these files, I'd recommend following the steps 
> outlined in:
>
> $ HELP/MESSAGE MULTALLOC
>

Isn't that what the OP said they were doing? I was only addressing the
question about working with files that have corrupted names. Many people do
not know how to work with files using only the FID. Any DCL commends that
work with files (DIRECTORY, COPY, DELETE, TYPE, RENAME, PRINT etc.) should
working using the FID syntax. If you know the directory the file is in then
you can prefix the FID with the directory and the command will work a lot
faster than going through the whole directory structure. I'm not sure if
using the FID as part of the filename is even documented, but if you have
filenames that are not valid then using the FID syntax is the best way to
work with them.

Peter Weaver
www.weaverconsulting.ca
Winner of the OpenVMS.org Readers' Choice Award for System
Management/Performance
http://www.linkedin.com/in/peterweaver