Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!husc6!cmcl2!brl-adm!adm!MANSFIEL%EMBL.BITNET@wiscvm.wisc.EDU
From: MANSFIEL%EMBL.BITNET@wiscvm.wisc.EDU (Niall Mansfield)
Newsgroups: comp.unix.wizards
Subject: dump - Sun/BSD dump-level idiosyncracies?
Message-ID: <8034@brl-adm.ARPA>
Date: Fri, 26-Jun-87 00:34:55 EDT
Article-I.D.: brl-adm.8034
Posted: Fri Jun 26 00:34:55 1987
Date-Received: Sat, 27-Jun-87 05:32:19 EDT
Sender: news@brl-adm.ARPA
Lines: 35


Sun say that if you dump at e.g. level 4, you get all files
which changed since the last level 0, 1, 2, or 3.  Thus, if
you keep on doing level 4's, your dump is getting bigger and
bigger, e.g. if the sequence is

files changed:     C1   C2   C3   C4   C5   C6   C7   ...
dump level done:      4    4    4    4    4    4    4   ...

where Ci is the set of files changed since the last dump (of
any level), then your successive dumps will contain

1st dump:       C1
2nd             C1 C2
3rd             C1 C2 C3
4th             C1 C2 C3 C4
etc

Now, Eric Foxley  (in his book "Unix for Super-Users") says
that  you dump repeatedly at level 4 you would get

1st dump:       C1
2nd             C2
3rd             C3
4th             C4
etc

which seems a more truly "incremental" method.

Is this a difference between Sun/BSD and the AT&T line?
What is the rationale behind it?

Thanks,

Niall