T.R | Title | User | Personal Name | Date | Lines |
---|
630.1 | Practicality | VROSRV::HEAFEY | | Wed Dec 16 1987 12:35 | 8 |
| I know this is a "close the barn door after the cow got out" answer
but, when the output volume is magtape, ALWAYS INITIALIZE the tape.
I've had the same experiences under 4.5 and 4.6. I realize that this
isn't a very "technical" solution but, it is practical. As an operator
my primary concern is being able to restore a user's files if the need
arises.
Dave H.
|
630.2 | | RDGENG::LESLIE | Andy, CSSE DECnet/OSI Program, WG5. | Wed Dec 16 1987 14:22 | 3 |
|
I DID initialise the !@#$%^ tape!! As far as I can see BACKUP OVERWROTE
THE LABEL.
|
630.3 | | RDGENG::LESLIE | Andy, CSSE DECnet/OSI Program, WG5. | Thu Dec 17 1987 07:34 | 9 |
|
Anyhow, I initialised the tape, MOUNTED it /BLOCK =8192 and when
I tried to copy the file I got
-RMS-E-CRE, ACP file create failed
-SYSTEM-F-BADATTRIB, bad attribute control list
%COPY-W-NOTCOPIED, DUA1:[TAPE]TAPE.A;1 not copied
More ideas please!
|
630.4 | | RDGENG::LESLIE | Andy, CSSE DECnet/OSI Program, WG5. | Thu Dec 17 1987 11:39 | 1 |
| Worked out why that happened - RMS only allows 2048 byte records.
|
630.5 | How I've done it before | OVDVAX::LENNIG | Dave, SWS, @CYO Cincinnati | Thu Dec 17 1987 16:34 | 25 |
| Andy, I assume you did your MOUnts /FOR (you didn't indicate so).
Reason I bring it up, is that if you didn't, the MTAACP will gobble
up the VOL and HDR records on the tape, and DUMP won't see them.
(At least that's the way I recall things.)
Regardless, I have hacked together tapes like this before, and have
a program (written in Fortran, slow as hell) that in conjunction
with certain DCL commands will allow you to recreate your tape.
The basic procedure I would use is...
Run program, which reads blocks off the tape and writes them to
disk, against the 'bad' tape. You now have a image of the tape.
Mount a scratch tape, init it, and backup an 'empty' disk, using
the same volume and saveset names as the prior attempt. Rewind
the tape (SET MAG/REW) space it forward to the beginning of the
first data block (SET MAG/???) and then run the program to copy
the image back onto the tape.
If you need a copy of the program (I got it from Z, who got it from
Y, who got it from X...) send mail. There's nothing particularly
fancy or unique about the program, it just issues QIO's to read
the tape and WRITEs to disk.
Good luck,
Dave
|
630.6 | I've been there | MAY20::MINOW | Je suis marxiste, tendance Groucho | Thu Dec 17 1987 21:15 | 6 |
| Andy, feel free to copy BOLT::DECUSC$LIBRARY:[MISC]READTAPE.C or
BOLT::USER:[MINOW.BIN]READTAPE.EXE.
READTAPE was written to read Unix TAR tapes onto disk.
Martin.
|
630.7 | Idle question | LESLIE::LESLIE | Andy, CSSE DECnet/OSI Program, WG5. | Fri Dec 18 1987 06:12 | 1 |
| Where is BACKUP format documented?
|
630.8 | | LESLIE::LESLIE | Andy, CSSE DECnet/OSI Program, WG5. | Fri Dec 18 1987 08:58 | 13 |
| interesting.
JC, a friend in need if ever there was one, has rebuilt the tape.
The two problems were
1) Backup overwrote the label
2) the 'saveset' was marked as having variable length records, which
meant that it wasn't a saveset, according to backup.
A little twiddling and all is well.
Thankfully.
|
630.9 | Where, oh where | MDVAX3::COAR | My hero? Vax Headroom, of course! | Fri Dec 18 1987 12:55 | 10 |
| Re .7:
> Where is BACKUP format documented?
In the 'fiche. It used to be in an .MDL file, but it's probably
SDL by now. Somewhere, I have a more-or-less complete copy of the
.MDL file (that I typed in by hand - oh groan!). Let me know if
you want it..
#ken :-)}
|
630.10 | | TUBORG::GROSSMAN | Expectation invites disappointment | Sun Dec 20 1987 19:01 | 8 |
| The best document for the backup saveset format was found in
appendix B of the V3 Utilities Reference Manual. To get an online
feel for it try BACKUP/LIST/ANALYZE. This breaks up the saveset
field by field and displays them.
- Russ -
|