Very Computer

I have several 2.7GB (after formatting) disks on my HP workstations
(one HP 735, one 715).  Our group's backup device is an Exabyte 8mm
(2.2 GB) drive,* off of a Sun Sparc 2.  To back up the HP's
disks, I'm forced to use rdump, or to nfs mount and use tar or cpio
on the Sparc 2.

"rdump"ing from the HP to the Sun doesn't seem to work.  The problem
is that the disk volume is bigger than the tape.  Upon reaching the
end of the tape, rdump returns the question:

  DUMP: NEEDS ATTENTION: Do you want to restart?: ("yes" or "no")

If I answer yes, it restarts the entire dump.  If I answer no,
it aborts the entire dump.  But what I *want* to do is put in
a new tape and continue the dump.  How do I do this?

If dump/rdump cannot be used to back up a file system that's
larger than the tape capacity, then what do I do?  Any suggestions?

Thanks.

--
Chris Metzler
Department of Physics, University of Michigan           313-764-4607 (office)
Randall Lab, 500 E. University                          313-996-9249 (home)

"As a child I understood how to give; I have forgotten this grace since I
have become civilized." - Chief Luther Standing Bear

Quote:>"rdump"ing from the HP to the Sun doesn't seem to work.  The problem
>is that the disk volume is bigger than the tape.  Upon reaching the
>end of the tape, rdump returns the question:

>  DUMP: NEEDS ATTENTION: Do you want to restart?: ("yes" or "no")

>If I answer yes, it restarts the entire dump.  If I answer no,
>it aborts the entire dump.  But what I *want* to do is put in
>a new tape and continue the dump.  How do I do this?

Assuming that this is the Sun 2.3 Gig Exabyte... (Exabyte 8200 class)

rdump 0dsf 54000 6000 tapehost:/dev/rst0 /dev/whatever

or

rdump 0audsf dumparchive 54000 6000 tapehost:/dev/rst0 /dev/whatever

should take care of you. (source: SunOS 4.1.3_U1B man page for dump)

--
                          -- Dave
   Office of the Dean, College of Arts & Sciences, Northwestern University

That's what I'm doing.  Those are exactly the parameters I use; I'm also
using a blocking factor of 126.  Still, somehow, when I reach end-of-media,
the HP generates the error.

???

--
Chris Metzler
Department of Physics, University of Michigan           313-764-4607 (office)
Randall Lab, 500 E. University                          313-996-9249 (home)

"As a child I understood how to give; I have forgotten this grace since I
have become civilized." - Chief Luther Standing Bear

What I would do (and actually do here) is to adjust your tape density
and length parameters to simulate whatever size tape drive you have
(remember, dump was written way back when 1600 BPI, 2300ft tapes were
the norm), and dump will pause after writing "X" bytes of data and ask
you to put in another tape.

I use the "b", "d", and "s" parameters on my dump command for maximum
performance:
        A "b" (blocking) setting of 128 works good over the net.
        A combination of "d" set to 1600, and "s" set to 67000
                simulates a 1.3GB DDS drive.

 Here's an example of how this works:

  DUMP: Date of this level 0 dump: Thu Dec 22 10:54:45 1994
  DUMP: Date of last level 0 dump: the epoch
  DUMP: Dumping /dev/rdsk/6s0 (/) to /dev/null
  DUMP: This is an HP long file name filesystem
  DUMP: mapping (Pass I) [regular files]
  DUMP: mapping (Pass II) [directories]
  DUMP: estimated 546527 tape blocks on 0.44 tape(s).

You can see that if I had about 1300000 tape blocks, it would report
about 1.0 tapes needed.

Just fiddling here, I see with a "s" setting of 100000:

  DUMP: Date of this level 0 dump: Thu Dec 22 11:15:24 1994
  DUMP: Date of last level 0 dump: the epoch
  DUMP: Dumping /dev/rdsk/6s0 (/) to /dev/null
  DUMP: This is an HP long file name filesystem
  DUMP: mapping (Pass I) [regular files]
  DUMP: mapping (Pass II) [directories]
  DUMP: estimated 546570 tape blocks on 0.29 tape(s).

That looks about like a 2.0GB tape to me...

And with "s" set to 200000:

  DUMP: Date of this level 0 dump: Thu Dec 22 11:17:32 1994
  DUMP: Date of last level 0 dump: the epoch
  DUMP: Dumping /dev/rdsk/6s0 (/) to /dev/null
  DUMP: This is an HP long file name filesystem
  DUMP: mapping (Pass I) [regular files]
  DUMP: mapping (Pass II) [directories]
  DUMP: estimated 546572 tape blocks on 0.15 tape(s).
  DUMP: dumping (Pass III) [directories]

Again, just foolin' around, I did this:

  DUMP: Date of this level 0 dump: Thu Dec 22 11:24:20 1994
  DUMP: Date of last level 0 dump: the epoch
  DUMP: Dumping /dev/rdsk/6s0 (/) to /dev/null
  DUMP: This is an HP long file name filesystem
  DUMP: mapping (Pass I) [regular files]
  DUMP: mapping (Pass II) [directories]
  DUMP: estimated 546858 tape blocks on 14.70 tape(s).
  DUMP: dumping (Pass III) [directories]
  DUMP: dumping (Pass IV) [regular files]
  DUMP: Tape rewinding
  DUMP: Change Tapes: Mount tape #2
  DUMP: NEEDS ATTENTION: Is the new tape mounted and ready to go?:
  ("yes" or "no")

Heres where you would rewind the tape (if you were using a no-rewind
device), eject the tape, mount a new one, and type "yes" to the prompt.

Hope this gets you goin' ...

Scott Larsen

--------------------------------------------------------------------------
"When they were supposed to ram us with the guns, they either swam
 away or put their snouts on our shoulders, very affectionately.  
 They were the worst at taking orders."
                - Richard Trout, former mammal trainer for the U.S. Navy,
                  on attempts to train dolphins to perform underwater
                  guard duty with snout-mounted .45-caliber guns.

Or, you can do the remote dump to a system, like the Sun Solaris 2
systems, that handle physical EOM properly and don't require the
'd' and 's' paramater hack to allow multi-tape dumps to work properly.

(PS Can HP-UX provide disk partitions for multiple file systems
per disk yet?.  One file system per disk is bad enough with 1 Gbyte
disks and I see 9 Gbyte disks comming into play now.)
--
    /
   /  Jerry,  UNIX SysAdmin,  +1 619 5873065
  /