From owner-freebsd-geom@FreeBSD.ORG Tue Nov 23 18:44:19 2004 Return-Path: Delivered-To: freebsd-geom@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id BE97016A4CE for ; Tue, 23 Nov 2004 18:44:19 +0000 (GMT) Received: from lara.cc.fer.hr (lara.cc.fer.hr [161.53.72.113]) by mx1.FreeBSD.org (Postfix) with ESMTP id 0A09643D5A for ; Tue, 23 Nov 2004 18:44:19 +0000 (GMT) (envelope-from ivoras@fer.hr) Received: from [127.0.0.1] (localhost.cc.fer.hr [127.0.0.1]) by lara.cc.fer.hr (8.13.1/8.13.1) with ESMTP id iANIiEQt011855 for ; Tue, 23 Nov 2004 19:44:15 +0100 (CET) (envelope-from ivoras@fer.hr) Message-ID: <41A384FE.1070002@fer.hr> Date: Tue, 23 Nov 2004 19:44:14 +0100 From: Ivan Voras User-Agent: Mozilla Thunderbird 0.9 (X11/20041111) X-Accept-Language: en-us, en MIME-Version: 1.0 To: freebsd-geom@freebsd.org X-Enigmail-Version: 0.86.1.0 X-Enigmail-Supports: pgp-inline, pgp-mime Content-Type: text/plain; charset=ISO-8859-2; format=flowed Content-Transfer-Encoding: 7bit Subject: Big geom_raid3 problems X-BeenThere: freebsd-geom@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: GEOM-specific discussions and implementations List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 23 Nov 2004 18:44:19 -0000 Im trying to use graid3 in production but I've encountered a big problem: I'm creating raid3 volume from three discs: ad4, ad5, ad6. All is well (and I mean everything - removing, inserting, rebuilding the raid device, using a filesystem on it) until I write BSD disk label on the raid volume (/dev/raid3/geri3). At that time, it seems to write partial labels to each of the data drives (like it should, presumably, since it splits requests), but FreeBSD recognizes it! e.g. /dev/ad4a, /dev/ad4b, etc. This by itself wouldn't be bad, but it seems to confuse the hell out of geom_raid3. Here are the messages that it writes on subsequent boot or loading: GEOM_RAID3: Device geri3 created (id=7). GEOM_RAID3: Device geri3: provider ad4 detected. GEOM_RAID3: Device geri3: provider ad5 detected. GEOM_RAID3: Device geri3: provider ad6 detected. GEOM_RAID3: Device geri3: provider ad4 activated. GEOM_RAID3: Device geri3: provider ad6 activated. GEOM_RAID3: Cannot update metadata on disk ad5 (error=1). GEOM_RAID3: Device geri3: provider ad5 activated. GEOM_RAID3: Device geri3: provider raid3/geri3 launched. GEOM_RAID3: Device geri3: provider ad5 disconnected. GEOM_RAID3: Device geri3: provider ad5 detected. GEOM_RAID3: Device geri3: provider ad5 activated. It's weird and disconcerting. Sometimes, the "cannot update metadata" message appears more than once. Usually that means the raid got completely botched (sometimes the computer hangs shortly afterwards): GEOM_RAID3: Device geri3 created (id=7). GEOM_RAID3: Device geri3: provider ad4 detected. GEOM_RAID3: Device geri3: provider ad5 detected. GEOM_RAID3: Device geri3: provider ad6 detected. GEOM_RAID3: Device geri3: provider ad4 activated. GEOM_RAID3: Device geri3: provider ad6 activated. GEOM_RAID3: Cannot update metadata on disk ad5 (error=1) GEOM_RAID3: Device geri3: provider ad5 activated. GEOM_RAID3: Cannot update metadata on disk ad5 (error=1) GEOM_RAID3: Device geri3: provider raid3/geri3 launched. GEOM_RAID3: Device geri3: provider ad5 diconnected. [hang right before "Mounting root" message should appear (root is on a different drive)] Curiosly, the second case appears only when I create the (ufs2) filesystems in /dev/raid3/geri[a,b,d] partitions. I really need a way to partition the device. I tried fdisk but it wants a physical device (with geometry) to work. Creating raid volume from partitions existing partitions instead of devices seems to confuse it also. This is on 5-STABLE from yesterday. Also, at one time, I accidentally wrote a disklabel to /dev/ad4 instead of the raid volume, and it suceeded, though geom_raid3 detected it and started synching the provider. Shouldn't the devices be locked or something?