On Fri, 7 Jan 2000, Michael Reifenberger wrote:
> Hi,
> after enabling the sym-driver I get drive-lockups after some time of accessing
> the disks hanging on the sym-driver.
> It seems that at least on disk hangs up (steady disk light) until a bus-reset
> "(noperiph:sym0:0:-1:-1): SCSI BUS reset detected" occurs.
>
> A difference between ncr and sym-driver is that the sym-driver probes the disks
> as (excerpt from dmesg_sym.txt):
>
> da1 at sym0 bus 0 target 1 lun 0
> da1: <IBM DDRS-39130D DC1B> Fixed Direct Access SCSI-2 device
> da1: 80.000MB/s transfers (40.000MHz, offset 15, 16bit), Tagged Queueing Enabled
> da1: 8715MB (17850000 512 byte sectors: 255H 63S/T 1111C)
>
> whereas the ncr-driver probes as (excerpt from dmesg_ncr.txt):
>
> da1 at ncr0 bus 0 target 1 lun 0
> da1: <IBM DDRS-39130D DC1B> Fixed Direct Access SCSI-2 device
> da1: 40.000MB/s transfers (20.000MHz, offset 15, 16bit), Tagged Queueing Enabled
> da1: 8715MB (17850000 512 byte sectors: 255H 63S/T 1111C)
>
> Usually the IBM-FAQ suspects a cabling/termination Problem if one has a Problem
> at 40MHZ.
Any SCSI related FAQ will suggest the same, in my opinion. But some SCSI
devices may have larger margin than some others regarding cabling /
termination problems and can be able to accomodate tiny problems. Some
other devices may just fail in same situations.
> But I have the 4 disks in a external case from kingston with a proper kable and
> terminator and furthermore LVD-drives should have a better signal quality.
> Furthermore I have zero problems at 20MHZ.
FYI, I have seen UltraWide SCSI Buses (single-ended) that continued work
reasonnably well with a terminator just missing (was intentionnal for
testing purpose). As a strangeness, only a single disk (had 4 on the BUS)
required to decrease speed at 33 MB/s (16.? MHz wide) for it to work
apparently (seems actually) well. As you can see, even a serious BUS
problem may, in some situations, allow the SCSI system to work by just
lowering speed of some devices on the bus.
> You can see in neg_before.txt the output of a 'camcontrol neg' which look right,
> in neg_after.txt the output of the same command after the disk hang up and
> the bus reseted.
> neg_ncr.txt contains the output using the ncr-driver.
>
> BTW: I have the same problems with the ncr-driver when using the kernel-config
> parameter: SCSI_NCR_DFLT_SYNC=10
>
> And now the question: Whats going wrong?
Probably something related to the SCSI BUS (or devices), but not to
software drivers, in my opinion.
> If it is the HW, how can it be proven? Are it the IBM-disks? Kabel?...
> Is it the Software?
> If I want to live with 40MB/s is there a knob to pre-set the speed in the
> kernel-config or at boot-time?
For the sym driver, you just have to configure SCSI devices accordingly in
the NVRAM. The sym driver hears from you by reading the NVRAM when it is
able to understand its layout and your kernel log messages let me think it
was able to read the controller NVRAM. (Ctrl/C at boot-up for the SYMBIOS
SDMS BIOS, and then configure your device settings).
> Any clues?
Regards,
Gérard.
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
