Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.3 alpha 4/15/85; site elsie.UUCP
Path: utzoo!linus!philabs!cmcl2!seismo!elsie!ado
From: ado@elsie.UUCP
Newsgroups: net.unix-wizards
Subject: cancel <2@elsie.UUCP>
Message-ID: <5123@elsie.UUCP>
Date: Thu, 16-May-85 11:07:03 EDT
Article-I.D.: elsie.5123
Posted: Thu May 16 11:07:03 1985
Date-Received: Fri, 17-May-85 05:47:18 EDT
References: <134@kontron.UUCP>
Control: cancel <2@elsie.UUCP>
Organization: NIH-LEC, Bethesda, MD
Lines: 19

The current way of handling file system dumps is to nail all processes using
the file system in question, unmount the file system, and then do the dump.
The nailing is done with the help of kludges like "fuser" and "ofile."

A thought, though:  it might be sufficient to simply "freeze" the file system
in question, rather than dismounting it.  You'd issue a "freeze" system call
which would write out any cached data and then set things up so that processes
that tried to write to the file system in question would block.  Then you'd
proceed with the dump.  When the dump was done, you'd issue a "thaw" system
call that would arrange things so that processes were once again allowed to
write to the file system in question, and would also unblock any processes
that got blocked by the "freeze."

Yes?  No?
--
UNIX is an AT&T Bell Laboratories trademark.
--
	UUCP: ..decvax!seismo!elsie!ado    ARPA: elsie!ado@seismo.ARPA
	DEC, VAX and Elsie are Digital Equipment and Borden trademarks