From owner-freebsd-questions Sun Oct 20 4:13:53 2002 Delivered-To: freebsd-questions@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 08A4D37B407 for ; Sun, 20 Oct 2002 04:13:51 -0700 (PDT) Received: from iota.root-servers.ch (iota.root-servers.ch [193.41.193.195]) by mx1.FreeBSD.org (Postfix) with SMTP id 677BE43E91 for ; Sun, 20 Oct 2002 04:13:49 -0700 (PDT) (envelope-from gaml@buz.ch) Received: (qmail 45585 invoked from network); 20 Oct 2002 11:13:37 -0000 Received: from dclient217-162-94-86.hispeed.ch (HELO gaxp1800.root.li) (217.162.94.86) by 0 with SMTP; 20 Oct 2002 11:13:37 -0000 Date: Sun, 20 Oct 2002 13:13:53 +0200 From: Gabriel Ambuehl X-Mailer: The Bat! (v1.60q) Educational Reply-To: Gabriel Ambuehl X-Priority: 3 (Normal) Message-ID: <5860268296.20021020131353@buz.ch> To: questions@freebsd.org Cc: sos@FreeBSD.org Subject: ATA(4) RAID support broken? Why won't it boot from degraded arrays? MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG Hi, I was experimenting with Highpoint 370 based IDE RAID cards today (BIOS rev. 1.0.3, flash tool from their site won't work so I can't upgrade it) and they will happily mirror the data to both disks in RAID 1 mode. But the fault resilience of the thing isn't what one would expect as the booting process will hang at the point / gets mounted from /dev/ar0s1a if one of the drives got disconnected before as it states it isn't able to mount the partition. The only IDE RAID cards which don't seem to exhibit this problem are the 3Ware Escalade series but those have their own share of trouble: as with the real RAID controllers, they will store partition information in NVRAM or some other rather weird place so if you take a disc that was used with an Escalade controller, you can't see any data on it when connected to a normal ATA controller. In case the partition data is saved on NVRAM, this is particularly bad if your controller dies but your drives are still living meaning that while you still have your data you can't access them because of lacking partition data which isn't exactly what is asked for IMHO. Note that 3ware controllers have their own twe driver and thus aren't really comparable to what happens with ata(4) driven cards such as Highpoint and Promise. I think it is ata(4)'s fault as it still could read the data from the remaining disk and thus continue booting. W2K for one can do that with degraded RAID 1 arrays so it can't be the hardware's fault. Furthermore, as booting will work up until it tries to actually mount partitions seems to be yet another pointer to the fact that there's a problem with the driver. regards, Gabriel To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message