[Info-vax] MOUNT and filesystems, was: Re: RMS record metadata, was: Re: Re; Spiralog, RMS Journaling (was

[Stephen Hoffman]

On 2016-06-27 17:39:55 +0000, Michael Moroney said:

> I accidentally produced something that would confound this.  I put an 
> ODS-5 VMS filesystem on a thumb drive.  Later I stuck the drive in a PC 
> and formatted it.  The PC then saw it as an (empty) FAT-32 drive.  I 
> don't think I put anything on the drive from the PC. Later I put it 
> back into a VMS system intending to init it with a VMS file system 
> again but I $ MOUNTED it with no problem.  All the necessary VMS bits 
> to mount it were still there. I did an $ ANALYZE/DISK expecting 
> something to be corrupted, yet it passed.   The VMS file system 
> structure was still there. Back to the PC and it was still FAT...

The FAT and ODS-5 structures overlap, so that disk was honked, and the 
tools involved were clearly not detecting the honkage; the 
inconsistencies.

> What should a file system agnostic MOUNT do with this?  I should try to 
> see if I can recreate it and see what is in block 0/the GPT blocks ...

Better detect a corrupt file system?   But for dual-format-capable file 
systems such as ODS-2 or ODS-5 and ISO-9660 systems, a switch or 
qualifier is necessary to mount a file system further down the search 
order from what's initially found, when the file systems are 
dual-format.  This qualifier already exists in OpenVMS.   There'll be 
yet added work here, if (when?) OpenVMS supports GPT disks and disk 
partitioning, too.

The sys$setboot tool can sniff some of the basic file system structures.

-- 
Pure Personal Opinion | HoffmanLabs LLC