From owner-freebsd-scsi Tue Sep 14 8:43:47 1999 Delivered-To: freebsd-scsi@freebsd.org Received: from arjun.niksun.com (gw.niksun.com [206.20.52.122]) by hub.freebsd.org (Postfix) with ESMTP id 6B38A15081 for ; Tue, 14 Sep 1999 08:43:38 -0700 (PDT) (envelope-from ath@niksun.com) Received: from stiegl.niksun.com (stiegl.niksun.com [10.0.0.44]) by arjun.niksun.com (8.8.8/8.8.8) with ESMTP id LAA16690; Tue, 14 Sep 1999 11:43:37 -0400 (EDT) Received: (from ath@localhost) by stiegl.niksun.com (8.9.2/8.8.7) id LAA06343; Tue, 14 Sep 1999 11:43:36 -0400 (EDT) (envelope-from ath) To: kachun@pathlink.com (Kachun Lee) Cc: freebsd-scsi@freebsd.org Subject: Re: data corruption when using aic7890 References: <199909082031.QAA74557@stiegl.niksun.com> <14295.54853.441698.273640@grasshopper.cs.duke.edu> <199909092117.OAA39115@pathlink.net> From: Andrew Heybey Date: 14 Sep 1999 11:43:36 -0400 In-Reply-To: kachun@pathlink.com's message of "Thu, 09 Sep 1999 14:17:03 -0700" Message-ID: <85btb5fqhj.fsf@stiegl.niksun.com> Lines: 50 X-Mailer: Gnus v5.5/XEmacs 20.4 - "Emerald" Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org >In article <14295.54853.441698.273640@grasshopper.cs.duke.edu>, you say... >> >> >>Andrew Heybey writes: >> > gallatin> What!?? You mean if I just move the cable to the 20Mhz >> > gallatin> connector, all my problems will just go away??? Even >> > gallatin> though its just a different port on the same PCI >> > gallatin> controller? (at least on my P2B-LS) >> > >> > Well, I don't know if *your* problems will go away (I'm not 100% sure >> >>There is still no joy in mudville -- I'm getting the errors even after >>moving the connector. >> >>I've upgraded my P2B-LS bios from 1006 to 1011. After the bios >>upgrade, the 7890 bios went from v2.01 to v2.11. While I had a >>monitor on it, I also took the PCI latency setting from 32->128. >>I'm re-running my tests now to see if the bios upgrade had any >>effect. >> >>Sigh, >> >>Drew > >I believe I am seeing similar problem as you did. I have this NFS server - >P2B-DS (only 1 CPU) + extra 7890 + 13 ST118273LW (18G LVD) - panics once every >several days with 'ufs_dirbad: bad dir' or 'fault on nofault entry'. I spend >past 4 months swapping out every pieces of hardware with our other systems that >would run for months. The only variable is that this is our 1st server using >LVD (Ultra2) drives. You may have confirmed my suspicion that there was a >problem with 7890 running Ultra2 drives. I tried upgrade the adaptec BIOS to >v2.11 and increaed the PCI lateny setting to 128, but that did not solve the >problem. > >I am at the point of giving up on the adaptec controller. I saw you mentioned >that you had no problem with the NCR 35C875 chip. I was wondering which board >manufacturer did you use . Do you know if they make a dual channel LVD >(Ultra2-wide - we need the LVD bus because of the cable length) board using the >NCR 35C875(6)? Or anyone has success with one? > >Regards I have tried a Tekram card using the NCR chip with LVD. It worked fine (didn't see the corruption problems), but for our purposes was slower than just plugging the disks into the Ultra (20MHz, 40MB/s) connector of the aic7890. However, if you need LVD for the longer cable lengths, maybe you should try the Tekram. It was model DC-390U2W. andrew To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message