Date: Mon, 7 Feb 2000 20:15:38 +0100 From: "Jakob Alvermark" <alvermark@teligent.se> To: "Kenneth D. Merry" <ken@kdm.org> Cc: <scsi@FreeBSD.ORG> Subject: Re: Problems with Adaptec card and FreeBSD 2.2.2 (fwd) Message-ID: <005101bf719f$b8a4a580$d14ea3c3@teligent.se> References: <20000207133426.A23334@teligent.se> <Pine.BSF.4.10.10002081058290.78201-100000@teligent.se> <20000208105150.A31083@panzer.kdm.org>
next in thread | previous in thread | raw e-mail | index | archive | help
----- Original Message -----
From: Kenneth D. Merry <ken@kdm.org>
To: <alvermark@teligent.se>
Cc: <scsi@FreeBSD.ORG>
Sent: den 8 februari 2000 18:51
Subject: Re: Problems with Adaptec card and FreeBSD 2.2.2 (fwd)
> On Tue, Feb 08, 2000 at 11:10:24 +0100, Jakob Alvermark wrote:
> > On Mon, 7 Feb 2000, Kenneth D. Merry wrote:
> > > The "timed out in {datain|dataout|command}" phase messages almost
always
> > > mean there is a cabling or termination problem on the bus. It means a
> > > signal got stuck on the bus somewhere.
> > >
> > > There were certainly bugs in the old SCSI layer that could get
triggered
> > > under heavy load, but this isn't one of them.
> > >
> > > These sorts of problems can be intermittent, so it isn't surprising
that
> > > you see them once a week under high load.
> > >
> > > The new Adaptec driver and the CAM SCSI layer handles problems like
this a
> > > little better, but there are still no guarantees when signals are
getting
> > > stuck on the bus.
> > >
> > > So check your cabling and termination.
> >
> > I have checked all the cabling and termination on one of the machines
> > (which was sent to me). Everything seems ok, there is only one disk,
which
> > is terminated, and one cable. The SCSI-card itself was set to automatic
> > termination, could that be a problem?
>
> It would only be a problem if the Adaptec driver isn't dealing with
> automatic termination properly in FreeBSD 2.2.2. That is certainly a
> possibility.
>
> > I was able to reproduce the problem by creating a number of processes
that
> > reads and writes a lot on the disk. After a couple of hours the machine
> > froze and showed the above message.
> >
> > I have now changed the termination on the card to "ON", and running the
> > same test.
>
> You might also want to check and see if the cable is too close to the
power
> supply. You may also want to just replace the cable with a new one and
see
> what happens.
>
> I've got a cable that looks okay, yet it caused sporadic (every few days)
> SCSI parity errors. Replacing the cable got rid of the errors.
Broken cables seems unlikely, it happens on 4 machines, I don't think all of
them could be broken. But the theory about closeness to the power supply is
interesting. In the machine i opened, the cable was actually strapped to the
power supply.
Maybe I should let the cable hang freely?
/Jakob
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?005101bf719f$b8a4a580$d14ea3c3>