home *** CD-ROM | disk | FTP | other *** search
Wrap
Received: from flibble.cs.su.oz.au (srn-ppp-epping.cs.su.OZ.AU [129.78.111.83]) by nacm.com (8.6.10/8.6.9) with ESMTP id OAA09713 for <executor@nacm.com>; Mon, 1 May 1995 14:07:31 -0700 Received: (from srn@localhost) by flibble.cs.su.oz.au (8.6.9/8.6.9) id HAA26497; Tue, 2 May 1995 07:06:18 +1000 From: Stephen Robert Norris <srn@flibble.cs.su.oz.au> Message-Id: <199505012106.HAA26497@flibble.cs.su.oz.au> Subject: Re: Possible way to improve Executor/Linux To: tjrc1@mole.bio.cam.ac.uk (Tim Cutts) Date: Tue, 2 May 1995 07:06:18 +1000 (EST) Cc: executor@nacm.com In-Reply-To: <9505010903.AA09434@mole.bio.cam.ac.uk> from "Tim Cutts" at May 1, 95 10:03:30 am X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 558 Sender: owner-executor@nacm.com Precedence: bulk > I have discovered that some filesystem problems with E/L 1.99k were > due to executor getting confused about the status of files, and that > this was particularly a problem with floppy access, and was solved by > manually sync'ing the filesystems in another window. So, I figure > that perhaps executor ought to call sync(2) after every filesystem > write. It would make it a bit slower, but might prevent a lot of > these problems. Or has this been done with 1.99m? (I haven't tried it > yet) > > Tim. fsync(2) would be a better choice. Stephen