From owner-freebsd-scsi@FreeBSD.ORG Sat Mar 17 21:50:57 2007 Return-Path: X-Original-To: freebsd-scsi@freebsd.org Delivered-To: freebsd-scsi@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 06D3416A406; Sat, 17 Mar 2007 21:50:57 +0000 (UTC) (envelope-from tom@samplonius.org) Received: from ly.sdf.com (ly.sdf.com [216.113.193.83]) by mx1.freebsd.org (Postfix) with ESMTP id 8FEF913C45D; Sat, 17 Mar 2007 21:50:56 +0000 (UTC) (envelope-from tom@samplonius.org) Received: from localhost (localhost [127.0.0.1]) by ly.sdf.com (Postfix) with ESMTP id BCCFA114447; Sat, 17 Mar 2007 14:29:44 -0700 (PDT) X-Virus-Scanned: amavisd-new at X-Spam-Score: -3.958 X-Spam-Level: X-Spam-Status: No, score=-3.958 tagged_above=-10 required=6.6 tests=[ALL_TRUSTED=-1.8, AWL=0.441, BAYES_00=-2.599] Received: from ly.sdf.com ([127.0.0.1]) by localhost (ly.sdf.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id N+TvOrHdLDxv; Sat, 17 Mar 2007 14:29:42 -0700 (PDT) Received: from ly.sdf.com (ly.sdf.com [216.113.193.83]) by ly.sdf.com (Postfix) with ESMTP id 9104D114446; Sat, 17 Mar 2007 14:29:42 -0700 (PDT) Message-ID: <8169444.541174166982500.JavaMail.root@ly.sdf.com> Date: Sat, 17 Mar 2007 14:29:42 -0700 (PDT) From: Tom Samplonius To: John Nielsen In-Reply-To: <200703161159.37735.lists@jnielsen.net> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit X-Originating-IP: [216.113.193.85] Cc: freebsd-scsi@freebsd.org, freebsd-hackers@freebsd.org Subject: Re: iSCSI boot mussings X-BeenThere: freebsd-scsi@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: SCSI subsystem List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 17 Mar 2007 21:50:57 -0000 ... > A truly standalone iSCSI client will most likely want to use a TOE > card, which > to the OS looks like any other SCSI adapter. (I'm unsure which if any > such > cards are currently supported in FreeBSD, but that's a tangential > question.) I really doubt this. TOE iSCSI cards are quite rare. Qlogic has one, and it could work under FreeBSD, as it looks similar the other Qlogic HBA cards. But they are still pretty rare. And TOE doesn't necessarily imply iSCSI. Various cards existing ethernet cards have various levels of offload support now. I think the general case is going to succeed: ethernet cards with acceleration, not iSCSI HBAs. > Machines with iSCSI-capable BIOS'es are an inbetween case. Allowing > such > machines to be standalone clients would require things like the > initiator > name, the initiator's IP address and netmask, the target's IP address, > and > the target's (volume) name to be hard-coded in the kernel. It would be > nice > to support this scenario, but IMO it's the one with the lowest > benefit/cost > ratio. Maybe, maybe not. Probably more people have iSCSI BIOSes in their servers now, than iSCSI dedicated HBAs. IBM introduced iSCSI boot code into many of their server BIOSes last year. I don't think it is that hard either. I don't think you will have to hard code into the kernel, but at least make sure the targets match what the BIOS has. It could just be a boot option, as the boot loader just uses BIOS calls, which would hit read from iSCSI. If you don't have an iSCSI BIOS in your server (which probably every Dell, IBM, and HP server will have standard in every server this year), and you don't have one of the very rare iSCSI dedicated HBAs, then PXE booting an iSCSI enabled kernel is probably the way to go. > Making it easy to integrate iSCSI into existing environments (diskless > or not) > is IMO the biggest hole in the current implementation (the missing rc > and > fstab bits I mentioned before). > > JN Tom