Remote backup strange behaviour

From: Nikola Milutinovic <Nikola.Milutinovic_at_ev.co.yu>
Date: Mon, 24 Apr 2000 15:29:26 +0200

Hi all.

I remember this matter was discussed, but I'm pressed for time. I'm trying to do
a remote system backup with vdump + dd.

This is the line used to do it:

Uprava:/# rsh -l root Twinpeaks.ev.co.yu "vdump -0uCf - /clones/root" | dd
bs=60k of=/dev/nrmt0h

It all goes well, but when I try to restore (vrestore -t) on the backup machine
(Uprava). I get nonsense output, the terminal goes heywire and I have to kill
it.

Next I tried an experiment:

Uprava:/# vdump -0NC /
...
Uprava:/# mt rewind
Uprava:/# vrestore -t

THIS GOES OK.

Uprava:/# vdump -0Cf - / | dd bs=60k of=/dev/nrmt0h
.....
566+22 records in
567+1 records out

Why is there a different number of blocks?

Then I try to read it:

Uprava:/# vrestore -t

FAILS. It complains about "incomplete number of blocks".
Then I try:

Uprava:/# dd bs=60k if=/dev/nrmt0h | vrestore -f - -t

THIS GOES OK.

So, what is the story? I remeber something vaguely about redirections, pipes and
vdump/vrestore + dd issue. The main question is why does network backup fail? I
tried with rsh:

Uprava:/# dd bs=60k if=/dev/nrmt0h | rsh -l root twinpeaks "vrestore -f - -t"

No luck.

Nix.
-- 
42??!!! 7 and half million years and all you can come up
  with is 42?
Received on Mon Apr 24 2000 - 13:25:31 NZST

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:40 NZDT