You have to set an option to get early warning detection-
ST_ENABLE_EARLYWARN in your kernel config file, or use
'mt' to set it. Otherwise you go to hard physical EOT.

This was a requirement forced on me.

-matt


On Sat, 26 Jul 2003, David Gilbert wrote:

> I've got a problem.  I've got an Ultra 5 with a SCSI/ethernet card
> that probes up as isp0 and hme1.  The machine has NetBSD 1.6.1
> installed on the internal IDE drive.
>
> Under heavy use, the hme1 has poor performance a) because it's half
> duples and b) because it's spitting out console messages.  I solved
> this probem by inserting an fxp0 into the machine.
>
> Now I have two tape drives:
>
> st0 at scsibus0 target 0 lun 0: <EXABYTE, EXB-89008E030203, V37f> SCSI2 1/sequential removable
> st0: density code 39, 1024-byte blocks, write-enabled
> st0: sync (100.0ns offset 12), 16-bit (20.000MB/s) transfers
> st1 at scsibus0 target 15 lun 0: <HP, C5683A, C908> SCSI2 1/sequential removable
> st1: density code 37, variable blocks, write-enabled
>
> I havn't tried st1 yet (doing so now), but writing to st0 to the end
> of the tape produces endless copies of the following message:
>
> st0(isp0:0:0:0):  Check Condition on CDB: 0x0a 00 00 08 00 00
>     SENSE KEY:  Volume Overflow
>                 EOM Detected
>    INFO FIELD:  2048
>  COMMAND INFO:  55 (0x37)
>      ASC/ASCQ:  End-Of-Partition/Medium Detected
>
> ... it would seem that the driver should just return EOM to the tar
> process... but that doesn't appear to be the case.
>
> Dave.
>
> --
> ============================================================================
> |David Gilbert, Independant Contractor.       | Two things can only be     |
> |Mail:       dave@daveg.ca                    |  equal if and only if they |
> |http://daveg.ca                              |   are precisely opposite.  |
> =========================================================GLO================
>