| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 307.1 | It is rumored that, ... | REGENT::MINOW | Martin Minow -- DECtalk Engineering | Tue Sep 09 1986 13:24 | 10 |
| When the 780 was first released, the pack that came with the
system was the one used by manufacturing during final test.
It contained a copy of the VMS development system pack. Customers
put it on the shelf and used one of their own packs for installation
testing.
They managed to keep this a secret from Dec for a few years!
Martin.
|
| 307.2 | | PASTIS::MONAHAN | | Tue Sep 09 1986 15:49 | 11 |
| For a competent disk scavenger a simple change like that suggested
in .0 is not too effective. For last year at European DECUS we used
a F.S. diagnostic to erase one pack on a system, and then did image
backups of this to others. I expect this year we might be able to
do shadow set catchups instead of image backups for some systems,
but some microvax systems can be a problem.
If anyone has any really good ideas I would welcome them since
we will need to do this again this year after DECville.
Dave
|
| 307.3 | qio writelbn | IOSG::BAILEY | | Wed Sep 10 1986 08:41 | 4 |
| Start at LBN 0 and write blank blocks up to disk max block
|
| 307.4 | not quite so easy.. | PASTIS::MONAHAN | | Mon Sep 15 1986 07:40 | 14 |
| For a non-system disk almost any simple system will work. For
the system disk it is a little more of a problem, since there are
a number of critical system files that may be accessed at almost
any time. If this happens the system will crash, and the disk will
not be rebootable, although it will still probably have many of
the files untouched.
A programme doing logical block QIOs would have first to build
up a table of block numbers that must not be touched. Something
that runs without VMS, such as an HSC utility is simpler and faster,
but not all microvaxen have HSCs.
It is almost Catch-22, like the memory clearing programme that
prints out the checksum of zero to prove that memory has been cleared.
|
| 307.5 | Maybe one of these will work ... | 52362::COWAN | Ken Cowan | Sun Sep 28 1986 17:32 | 24 |
| At first I thought ANAL/MEDIA/EXERCISE would be useful, but
that doesn't quite work on the system device while the system
is running.
If you are in to hacking, it might be possible to get ANAL/
MEDIA to run standalone. Off hand, I'm not sure what needs
to be done.
An alternative is to fill up an appropriate sized disk with
trash, do am image backup and then use standalone backup to
write to the system disk. Note that getting standalone backup
to initialize the disk is a marginal solution. After the
disk is initialized, it can be recovered [with appropriate
hacking.]
Maybe use the FS diagnostic to reformat the disk?
Just thinking ...
KC
|
| 307.6 | IO$_WRITELBLK kinda worked for me | MDVAX3::COAR | A wretched hive of bugs and flamers. | Tue Nov 24 1987 01:43 | 15 |
| I wrote a program to do this when I had to return a system pack
to the site we had borrowed it from. The program started at LBN
<maxblock>, and zeroed it, then n-1, n-2, and so on. It printed
out little status reports every 100 blocks; after a while, I got
concerned, since it was running along merrily and the system remained
up, so I typed CTRL/Y - and the system crashed and remained unbootable.
Apparently, it had already gotten to the point of something that
would get invoked at image rundown. Oh, yeah - it did a $LCKPAG
to lock the entire working set into physical memory. Unfortunately,
the program was on the disk being wiped - so I lost it. I've sometimes
wondered what would have happened if I had left it alone... ;-)
Maybe I'll try it again sometime - `Coming to a System Near YOU!'
This was back under VMS Version 3.n.
#ken :-)}
|