From owner-freebsd-scsi Mon Dec 13 2:11:47 1999 Delivered-To: freebsd-scsi@freebsd.org Received: from mass.cdrom.com (castles516.castles.com [208.214.165.80]) by hub.freebsd.org (Postfix) with ESMTP id 9396B14FF7; Mon, 13 Dec 1999 02:11:43 -0800 (PST) (envelope-from msmith@mass.cdrom.com) Received: from mass.cdrom.com (localhost [127.0.0.1]) by mass.cdrom.com (8.9.3/8.9.3) with ESMTP id CAA21782; Mon, 13 Dec 1999 02:14:24 -0800 (PST) (envelope-from msmith@mass.cdrom.com) Message-Id: <199912131014.CAA21782@mass.cdrom.com> X-Mailer: exmh version 2.1.1 10/15/1999 To: "Oliver Blasnik" Cc: freebsd-scsi@FreeBSD.ORG, "Mike Smith" Subject: Re: Again: CRD-Raid-Controller and FreeBSD 3.x In-reply-to: Your message of "Mon, 13 Dec 1999 11:07:21 +0100." <000c01bf4551$d8862240$da1940c2@omnilink.de> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Mime-Version: 1.0 Content-Transfer-Encoding: quoted-printable Date: Mon, 13 Dec 1999 02:14:23 -0800 From: Mike Smith Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > Mike Smith wrote: > = > > No. FreeBSD starts assuming that a drive will handle 64 tags, but it= = > > expects that the drive will correctly report a queue-full condition s= o = > > that it can dynamically adjust this number downwards. > = > And this could not work as far these maximum of 32 commands is host-bas= ed, not lun-based. That would depend on the configuration of the unit. It'd work well for = the single-drive case. > > This may solve the "problem", but it will substantially degrade = > > performance in the case where there's only one array on the controlle= r. > Right. But on the other hand enables tq without bothering on system-cra= shed :) > Better slow than not running. Perhaps. Better fixed firmware than terrible performance. > > > > You might want to consider using a PCI:SCSI RAID controller like a My= lex = > > DAC960 or AMI MegaRAID. The host:cache bandwidth is _much_ better on= = > > these units, and they typically offer all of the features of the exte= rnal = > > units at a lower price. > > > = > *g* Not a possibility for me. > = > It has to be external for different reasons. We do not only have > FreeBSD, there's also Solaris and WinNT. NT supports these controllers, as does Intel Solaris. They probably have= = drivers for the PCI Sparc systems as well. > Connecting an external system > enables transparency. Some CRD's are connected to two machines, sometim= es > to share the array, sometimes to enable high-availability for a system > (hot-take-over of the drives). You could set up two CRD's to the same > drive bay to enable renundancy and cut off this single point of failure= =2E =2E.. if they worked properly. 8) > Last but not least, if one machine burns down, just take a new hardware= , > plug it onto the raid and switch it on - running and up again. Tell me = how > to do that with your controllers :) Easy. They all save their config on the array as well as in NVRAM. With= = the newer models you could even pull the battery-backed RAM module off = the burnt controller and save the cached write data as well. -- = \\ Give a man a fish, and you feed him for a day. \\ Mike Smith \\ Tell him he should learn how to fish himself, \\ msmith@freebsd.org \\ and he'll hate you for a lifetime. \\ msmith@cdrom.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message