[Info-vax] ODS-5 data/file recovery

[MG]

On 17-feb-2013 18:40, Hein RMS van den Heuvel wrote:
> I just downloaded florian V1.3-1 and checked the code. It is as I suspect
> :
>        printf("*** not a Files-11 Structure Level 2 disk volume\n");
>        SYS$EXIT(SS$_BADPARAM | 0x10000000);
> :
> Just drop the exit and see how far it gets.

That I will definitely try.  Any results will probably take a while, as
I'm following up your previous advice, to make a proper 'physical' back-
up (in combination with LDD).  It will take a while, as the volume is
over


> Having glanced over the Florian code it does pretty much what I expected it
> to do. I'm sure DFU does something very similar, possibly more robust/complete
> (no anal/disk/repair required, but suggested. Give it a go, but I expected the
> same result.

Thanks for having looked into it.  I wonder, too, I can't wait to try.


> Your goal is to find the file header for the deleted file back.
> The assumption is that the filename and the mapping pointers are untouched.
> So make the fileheader valid again, and the file is back, just not in a directory
> which is easy to fix (ANAL/DISK/REPAIR or my 'enter' tool below).

The file was placed in the root ([0,0]), so that should hopefully
simply things?  (Thanks for the fix either way though!)


> But since DFU did not find it, it is likely gone.

I fear so also, but, these are last ditch attempts and tries and
after that, I will give up and accept my losses.  It's not too
dramatic, fortunately.  The thing I miss most are some of my
administrative files.


> This is to be expected because VMS returns a deleted file header and
> blocks into cached structures to optimize the creation of new file,
> re-using what was just freed. I suspect it even goes to the begin
> (LRU) of the cache not the end, for immediate re-use. Yikes for you.

That's a very unfortunate mechanism indeed...


> I share Hoff's experience that folks are will to keep you busy for hours
> talking about possibly recovering data and then nothing happens (money).

I've never had experience with it, but that sounds like more
headaches.  I wonder also how many data recovery services out
there truly would know how to recover VMS data.


> You may be the good exception. I suspect there is at least a nice RX2620
> available as reward, once everything is proven stable huh?  !? :-).

You mean as a future fix for the caching or for recovering my
data?  Especially for the latter, I'd be open for a lucrative
deal indeed.


> [enter.c below]

Thanks, that looks excellent, I will build it later this evening
and let you know how it went.

  - MG