Date: 21 Mar 2001 13:26:09 +0200 From: Jussi Reissell <reissell@cc.helsinki.fi> To: Gérard Roudier <groudier@club-internet.fr> Cc: Jussi Reissell <reissell@cc.helsinki.fi>, freebsd-scsi@FreeBSD.ORG Subject: Re: Booting problem Message-ID: <87k85jmkha.fsf@mursu.pesa.fi> In-Reply-To: Gérard Roudier's message of "Tue, 20 Mar 2001 20:04:41 %2B0100 (CET)" References: <Pine.LNX.4.10.10103201930480.1462-100000@linux.local>
next in thread | previous in thread | raw e-mail | index | archive | help
Gérard Roudier <groudier@club-internet.fr> writes: > On 20 Mar 2001, Jussi Reissell wrote: > > > Sorry folks, this probably isn't a FBSD topic as such, more of a > > generic hw problem, I guess ... but maybe some tekram owner is able to > > give me advice. > > I never heared of this problem, neither am a Tekram owner. > So, I shouldn't have replied to your mail. :) Ok, I'll accept the disclaimer:) > I would suggest you to report the problem to Tekram support, and why not > upgrade the BIOS of the board if some new version is available. Alright. This didn't cross my mind. I'll start looking at all my BIOSes. > > I'm having a bit of a problem booting from an IBM disk hooked up to a > > Tekram 390u2w controller. At boot-time the controller refuses to load > > the BIOS and hence boot from the SCSI disk. The 390 spits out a: "No > > BIOS disk found! SCSI BIOS not installed" > > At least by default, the controller BIOS scans SCSI devices by ascending > SCSI ID number. If it is some timing problem, using a different SCSI ID > for your disk can make difference. It is just guessing... > > If, for example, your disk is configured for a very low SCSI ID, I would > suggest you to configure it for some higher SCSI ID value and force the > BIOS to scan some (non existing) SCSI IDs before this one. > Each SCSI ID scanned is about 250 ms elapsed. > I haven't any better idea. May-be you already tried it. There's so many BIOSes around that I get confused about who scans and what! With the Tekram card there's two different behaviour that I can actually see: 1) Occasionally things work as expected. Right after the regular BIOS, the Tekram SCSI BIOS kicks in and starts enumerating all the SCSI devices. After that, I get the normal responses from the boot blocks and I can boot from either an IDE disk or the IBM SCSI disk. 2) Most of the time, however the Tekram BIOS never shows up. Instead, I get the "NO BIOS disk found ..." message above. When this happens, I don't see the BIOS scanning anything and I'm only able to boot from IDE. I've switched around other ids but not the hard disk's. I'll try this yet. > > According to the Tekram manual, this is normal if only a CD-ROM is > > installed. I get this behaviour with only the IBM disk on board. > > > > The strange thing is that the controller occasionally is able to find > > the disk allright and boot from it but most of the time is not. > > > > What should I look for? I took out all the other SCSI devices, so as > > far as I can tell, the problem should be in the controller, cabling or > > the disk itself. Maybe termination? > > If the SCSI hard disk is working reliably (you may exercise the drive > after boot from IDE), then the SCSI BUS is likely correct. FBSD works extremely well with the whole SCSI chain. The sym driver has never produced any complaints at all and the IBM disk has been reliable as a data disk. No problems whatsoever. Even when at boot-up the BIOSes don't appear to play ball. I've done some elementary 'dd' testing and the SCSI drive comes up slightly slower than the IDE, but this most likely reflects the way I've done the testing, not some actual raw performance. My problem is that I simply don't know how for instance cabling and termination problems manifestate themselves. I'm I likely to see complaints from the driver first? Or could these go unnoticed to the driver and come up as possibly obscure performance problems or some such which would at least seem to be dependent on the phase of the moon? Just like my boot time problem appears to be. Settling this would eliminate cabling as one problem source and indeed would leave the BIOSes as the most likely candidate. Thanks a lot for your help. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?87k85jmkha.fsf>