Very Computer

Hi,

I have a:

Pentium 150 (Triton chipset)
UltraStor 14F ISA SCSI card
Conner 340MB SCSI HD
340MB IDE HD
Matsushita SCSI CDROM

and I recently have added a DEC 1.3GB SCSI HD.  However when trying to copy
a CD onto this new disk (viz. cp -Ra /cdrom /expansion) it crashes the machine
after about 30MB of writing.  On the console SCSI messages start scrolling
by and shortly after the whole thing is locked, with the SCSI light stuck on
and only the reset button seems to get a response.  I have tried using both
the UltraStor drivers that come with the kernel (one at a time of course) but
the only difference is when I reboot---the second driver spews a whole lot
of Ux14 message up the screen, as if it got congested and is now releasing
it all.

So I am unsure if this is a problem with the disk or the SCSI driver(s) or
the kernel or what (it happens the same under 1.2.13 and 1.99.9 and 1.99.10).
The disk has been used in an alpha previously with no problems so I think it's
a linux problem.  It's really annoying, I'd love to be able to use this new
disk!

Is this a known problem with this combo of hardware?  Please help?!

Cheers,

--
8<--------------------------------------->8

8<--------------------------------------->8

: I have a:
:
: Pentium 150 (Triton chipset)
: UltraStor 14F ISA SCSI card
: Conner 340MB SCSI HD
: 340MB IDE HD
: Matsushita SCSI CDROM
:
: and I recently have added a DEC 1.3GB SCSI HD.  However when trying to copy
: a CD onto this new disk (viz. cp -Ra /cdrom /expansion) it crashes the machine
: after about 30MB of writing.  On the console SCSI messages start scrolling
: by and shortly after the whole thing is locked, with the SCSI light stuck on
: and only the reset button seems to get a response.  I have tried using both
: the UltraStor drivers that come with the kernel (one at a time of course) but
: the only difference is when I reboot---the second driver spews a whole lot
: of Ux14 message up the screen, as if it got congested and is now releasing
: it all.

Further to this, I am now pretty sure it's reading from the CD that is the
problem---though only on large reads (> 50MB say).  I have done the following
tests:

         SCSI CD    to  external SCSI disk   failed (i.e hung machine!)
         SCSI CD    to  internal IDE  disk   failed (i.e. ditto)
internal SCSI disk  to  internal IDE  disk   success
external SCSI disk  to  internal IDE  disk   success
internal SCSI disk  to  external SCSI disk   success

so it's only the CD that's the problem.  In the kernel config I enabled both
SCSI disk and CD support so is this experienced by others?

--
8<--------------------------------------->8

8<--------------------------------------->8

: :
: : Pentium 150 (Triton chipset)
: : UltraStor 14F ISA SCSI card
: : Conner 340MB SCSI HD  SCSI id 0
: : 340MB IDE HD
: : Matsushita SCSI CDROM CR-800A SCSI id 6
: : DEC RZ58 1.3 GB SCSI HD SCSI id 4
:
: Further to this, I am now pretty sure it's reading from the CD that is the
: problem---though only on large reads (> 50MB say).  I have done the following
: tests:
:
:          SCSI CD    to  external SCSI disk   failed (i.e hung machine!)
:          SCSI CD    to  internal IDE  disk   failed (i.e. ditto)
: internal SCSI disk  to  internal IDE  disk   success
: external SCSI disk  to  internal IDE  disk   success
: internal SCSI disk  to  external SCSI disk   success
:
: so it's only the CD that's the problem.  In the kernel config I enabled both
: SCSI disk and CD support so is this experienced by others?

I have just upgraded to 1.99.12 and the problem persists. In fact it seems
to die earlier now (after 17MB rather than 30MB).  I have managed to write
down some of the message spat to the console as it dies:

scsi: aborting command due to timeout: pid 7143, scsi 0, channel 0, id 6,
      lun 0 0x08 03 62 9c 08 00
U14F0: abort, mbox13, target 6, pid 7143
U14F0: abort, mbox is in use.
U14F0: reset,.....

and many more 'reset' commands.  Any help to anyone in solving this one?

TIA,
--
8<--------------------------------------->8

8<--------------------------------------->8

Quote: Shepherd) writes:

I have an Ultrastor 34f (the VLB version of the same card) and looking thru
my manual for it, I might suggest going over a) the SCSI ID's of all devices,
especially the CD-ROM, b) checking out the 'synch negotiation' setting on
the 14f itself;  it almost looks like the CD-ROM is firehosing the buffer on
the board.  But then again I may be totally misunderstanding your trouble.  
I hope This at least points you in the right direction.

:
: I have an Ultrastor 34f (the VLB version of the same card) and looking thru
: my manual for it, I might suggest going over a) the SCSI ID's of all devices,
: especially the CD-ROM, b) checking out the 'synch negotiation' setting on
: the 14f itself;  it almost looks like the CD-ROM is firehosing the buffer on
: the board.  But then again I may be totally misunderstanding your trouble.  
: I hope This at least points you in the right direction.

Thanx for the followup.  I have checked the SCSI id's and there's no conflicts.
The synch negotiation I know nothing about however.  All I know is this machine
was working well under Win NT before I installed linux so I guess we can rule
out hardware faults.

I have just upgraded to kernel 2.0.0 and tried the CD copy test and got through
57 MB before the old troubles struck, which is an improvement I guess.

Occasionally the SCSI troubles start up when no CD activity is going, i.e.
just after a couple of days of regular disk activity and the machine doesn't
actually lock up but SCSI errors start scrolling up the screen (console
actually) and the only solution is the reset button.  The messages are
complaining about mboxes being locked, and it cycles through them from 0 or 1
to 15.  Also in the messages are reference to a pid that does not exist (in
fact it would be out of range on linux I think viz. 80000-ish).

Does this help at all anybody?

--
8<--------------------------------------->8

8<--------------------------------------->8

I hope this helps
    Martin