Date: Wed, 12 Aug 1998 13:04:06 +0200 From: Miroslav Kes <mira@rockwell.cz> To: "freebsd-scsi@FreeBSD.ORG" <freebsd-scsi@FreeBSD.ORG> Subject: Re: Heavy CAM Problems... Message-ID: <35D176A6.567EB43C@rockwell.cz> References: <199808111906.NAA06851@narnia.plutotech.com> <35D13957.D52AC90D@pubnix.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Alain Hebert wrote: > SOLVED! (Or at least isolated...) > > That's what is happening... > > The new cables (U2W: The one with a terminator and with all the wire > independently connecter to the connector (but still in serial fashion) ) > are kinda having quite a ride on EMI... Yeap, your PC emit like a small > Radio station! So you have to make sure where you route the cable > throughout the casing. > > Like, Justin said: that the Adaptec Guys thinks it hardware... Some > timing or access sequence were surely changed between 0520 and 0712 and > made that problem appear more frequently. > > So a good SCSI hygiene and good posture (don't route the cable around > the power, but near the casing following a single direction, not to > pinch or apply stress to the cable or wire...) might generally solve the > problem. > > (This can explain why some got 0x17d and others 0x17e... I was > re-arranging it tonight when I start getting 0x2!) Well, before the discussion on this topic started I thought it was my fault and I checked IRQ conflicts, improper termination and also routing of SCSI cables inside the chassis with respect to the power unit. I have already mentioned it in one of earlier messages. The symptom looked like as long as I had the machine on the table without the cover it worked and it stopped working when I put the cover on and put the machine on the floor. In the meantime I found one problem with termination and after solving it the problem disappeared and I thought I won. Three days later I had to move the computer to another place in our office so I had to stop it and boot again and the problem arose again. Playing with cable routing sometimes helped (for one or two boots). Now I cannot boot it at all. Any of tricks I used before doesn't work any more. The 2940U2W controller has three connectors Ultra2 68 pin, Ultra 68 pin and 50 pin. I'm using all of them so there are three cables with some devices attached. Today I inserted an IDE disk with 0520 snapshot and booted from it. As long as there is no cable with an device attached to the SCSI controller it boots without problems. If I attach ANY device no matter whether it is Ultra or Ultra2 type the boot doesn't complete (BIOS correctly reports which device is attached). I want to point out I have similar experience as you do but I haven't found any solution that would guarantee it was solved forever. Since the computer I'm talking about is intended to be our main fileserver and I need rock solid solution (that's why I have chosen FreeBSD). I would say the problem is still open. Going to play with memory timing. Mira -- ----------------------------------------------------------- | Miroslav Kes | |---------------------------------------------------------| | Rockwell Automation Ltd. | tel.: (+420) 2 2425 6913 | | Research Center Prague | fax: (+420) 2 250467 | | Americka 22 | e-mail: mira@rockwell.cz | | 120 00 Praha 2 - Vinohrady | | | Czech Republic | | ----------------------------------------------------------- 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?35D176A6.567EB43C>