From owner-freebsd-stable Wed Oct 10 16: 8:57 2001 Delivered-To: freebsd-stable@freebsd.org Received: from phoenix.shells.co.uk (phoenix.shells.co.uk [217.33.11.129]) by hub.freebsd.org (Postfix) with ESMTP id 7B97737B406 for ; Wed, 10 Oct 2001 16:08:47 -0700 (PDT) Received: from localhost (localhost [127.0.0.1]) by phoenix.shells.co.uk (Postfix) with ESMTP id 5ACF63822 for ; Thu, 11 Oct 2001 00:08:25 +0100 (BST) Date: Thu, 11 Oct 2001 00:08:25 +0100 (BST) From: Darkcyde X-Sender: jk@phoenix.shells.co.uk To: stable@freebsd.org Subject: 4.4 weirdness, dirprefs/RAID problem? Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG Hi there, I've been scratching my head over the last couple of days trying to get my head around a problem I've run into after getting 4.4 running on a machine. I have an Abit kt7a-raid board which has one of these evil raid-in-bios affairs using a Highpoint HPT370 controller, now before you scream that IDE raid is inherently b0rked or anything like that, I'd like to elaborate a little, heh. The bottom line is that the system ran perfectly fine with 4.3-STABLE, whereas with 4.4 it will die under certain conditions. We've been using swish-e for some database indexing which has quite a hefty performance footprint while it churns, to the point that it makes the box die under 4.4, something that didn't previously happen. I've tried in vain to track the source tree back to RELENG_4_3 and rebuild the whole lot, but upon attempting to build the world it fails during the bootstrapping procedure and refuses to continue on its merry way. I did however manage to succeed with building a 4.3-RELEASE-p20 kernel with the existing config I have, which boots fine, and upon testing out the indexing there are no errors and the box doesn't hang.. Whilst scoping about for any major changes with 4.4, I happened upon a post where someone (Kris?) mentioned that the new dirprefs code was not toggleable because it has no adverse effects hence there was no point in disabling it as it can only do good.. well, the states listed in top upon the box hanging are "FFS no" for swish-e itself, along with vfs_ type things for some of the other processes. I'll post what it looks like when I'm next physically at the machine but for the time being I'm reluctant to down it. Any help appreciated. Regards, J. -- Darkcyde (jk@dac.org) To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message