From owner-freebsd-scsi Thu Apr 24 08:11:46 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.5/8.8.5) id IAA10742 for freebsd-scsi-outgoing; Thu, 24 Apr 1997 08:11:46 -0700 (PDT) Received: from indigo.ie (aoife.indigo.ie [194.125.133.9]) by hub.freebsd.org (8.8.5/8.8.5) with ESMTP id IAA10737 for ; Thu, 24 Apr 1997 08:11:42 -0700 (PDT) Received: from indigo.ie (localhost [127.0.0.1]) by indigo.ie (8.8.5/8.8.5/INDIGO-HUB) with ESMTP id QAA05458; Thu, 24 Apr 1997 16:11:25 +0100 (BST) Message-Id: <199704241511.QAA05458@indigo.ie> To: "Justin T. Gibbs" Cc: freebsd-scsi@freebsd.org, judgea@indigo.ie, wilko@yedi.iaf.nl Subject: Re: Vendor specific ASCQ SCSI errors in 2.2-STABLE In-reply-to: Message from "Justin T. Gibbs" dated Monday at 10:10. From: Alan Judge Date: Thu, 24 Apr 1997 16:11:25 +0100 Sender: owner-freebsd-scsi@freebsd.org X-Loop: FreeBSD.org Precedence: bulk [Replying to a few messages at once.] Justin> Definitely trim that cable. Just be sure to clean the wires Justin> after you make the cut as most are braided and tend to Justin> fray/short the other lines after you cut them. I delayed replying until I saw what shortening the cable did. The system made nearly two days before producing that odd error. I guess I'll have to try canister swapping. >>>>> Wilko Bulte writes: Wilko> Any chance you can try things without the canisters? A lot of work alright, and I'm not sure I want to do that on a production machine. Bear in mind that I've never seen the same problem on the other bus (same config --- 4 drives in canisters). Both busses are pretty busy (two CCD stripes). I guess it might be a broken canister or a dodgy ribbon cable. I'll have to do some swapping, since I can't think of any better way to narrow down which canister. What's the likelihood that the disk producing the error is in the faulty canister? Wilko> An active terminator on the end of the bus instead of a Wilko> drive-internal terminator in a canister is also worthwile. Couldn't easily get ribbon cables with the correct number of connectors for this. In any case, the Quantum Atlas II claims to do active termination (controlled by a single jumper rather than a restistor pack). Wilko> How much stublength does each canister introduce? Quiet a bit. Maybe 20cm round trip. Wilko> Are these canisters designed with Ultra SCSI speeds in mind? The booklet doesn't mention Ultra, so I'd guess no. -- Alan Judge Phone: +353-1-6046901 Indigo Internet Services Fax: +353-1-6046948