Fixing Corruption on a VMS Disk

[rjarratt]

I have acquired an RF71 disk. Using the DUMP command it looks like it could be a VMS volume for VMS 6.0. However it won't mount because it can't find the home block. I tried to see if I could use the ANALYZE command to repair it, but that wants to be able to mount the disk first. Are there any tools to allow this disk to be analysed and maybe repaired so that it is mountable?

 

[daver2]

Just thinking aloud here - is there the equivalent in VMS of RSX11 to mounting the disk foreign? The disk is mounted - and it can then be 'manipulated' at the low level.

Dave

 

[bobaboba]

That would be mount/foreign of course.
But aside from backups I don’t know any VMS tools to manipulate the mounted/foreign volume.

 

[DDS]

No equivalent of chkdsk or fsck?

Eww!

Worst case would be using a sector editor and learning more about VMS file structures than you ever wanted to know.

 

[bobaboba]

.....but that wants to be able to mount the disk first. Are there any tools to allow this disk to be analysed and maybe repaired so that it is mountable?

I guess you can mount it with mount/foreign/nomount_verification? Not sure though and I don't know what you'd do after that.

 

[leegleason]

If it doesn't mount due to not finding a home block, it may be massively messed up, since ODS-2 includes a backup of it further along on the disk. You should dump out the default home bloke location (LBN 1) and see what it looks like....also try MOUNT/OVERRIDE-LIMITED_SEARCH (tells it to look really hard for a valid home block).

 

[rjarratt]

Not come across the MOUNT/OVERRIDE-LIMITED-SEARCH, but I was using VMS 5.4 so maybe its in a later version. I did try mounting it foreign, but the ANALYZE utility complains (which seems to defeat the object!). Understanding the file structure could be an option I suppose.

 

[leegleason]

Ah, likely it did come along later. Well, dump out the first block or two and post it here....I'd be interested to see if it looks anything like a home block or...something else. I recall having problems long ago with a system disk that kept getting corrupted. 3rd time it happened, I dumped the first block out, and found that it had "CPTSPL Exiting" written on it, along with a bunch of nulls. A privileged programmer had mixed up what channel he was using for messages, and was using QIOs with IO.WLB to write them...which, on a disk, went on the first LBN of the disk.

 

[rjarratt]

I got the disk out again to try some of the suggestions. The mount command went like this:

TEMP$ mount/over=(id,limited_search) $1$dia0
%MOUNT-W-ALHOMINV, invalid alternate home block on _$1$DIA0:
%MOUNT-W-QUOTAFAIL, failed to activate quota file; volume locked
-SYSTEM-W-BADIRECTORY, bad directory file format
%MOUNT-F-BADSECSYS, failed to create or access SECURITY.SYS
-SYSTEM-W-BADIRECTORY, bad directory file format
TEMP$

So I mounted it foreign again and dumped the first few blocks. I don't know how to read the file structure (yet). The dump is attached. I dumped the same blocks from a known good RF72. First glance suggests the bad disk has a home block but only one and perhaps shifted by a block. The good one may have multiple copies of the home block.

 

[Bitly]

Some reference materials for you:

VMS File System Internals
ODS2 file system reader (tested on Linux recently, MSDOS years ago, believed to work on VMS also)

CW