SUMMARY: Seagate Barracuda on Sparc 20 Sol 2.3

From: Mr T. Crummey (tom@sos.bangor.ac.uk)
Date: Tue Dec 13 1994 - 03:31:49 CST


Hello All,

Thanks to all below for their contributions. The problem was solved by getting
a new disk with a newer firmware revision. The original revision was 0014 but
version 0017 was the one that worked. My vendor swapped the disk and also
Seagate offered an upgrade.

Other solutions suggested:
Try shorter cables and active terminators
(both of which I'd tried before posting originally).

Try to set scsi_options=0x58 in /etc/system which sets
disk transfers to asynchronous which can decrease performance.

Check that the on disk terminator was disabled and that it is a unique SCSI ID.

Use fewer external devices. (I only had one on the machine that didn't work,
4 on the machine that did!)

One respondent said that Barracudas were differential SCSI devices. This is
only true if that is the device you ordered. They come in differential, wide
and wide differential forms as well as the narrow single ended I had.
It is true that you cannot mix these different disk types on the same bus.

I was also told to check out patch 101378 which fixes many SCSI problems in
Solaris 2. I had installed that patch also.

One person said that solaris cannot cope with partitions greater that 2 GBytes.
I believe this only true for Solaris 1.x.x and the limit is alot higher in
Solaris 2.x In any case, my biggest partition was 1.8GBytes.

Thanks again to the following and especially to Sean Boran, Stefan Eisenhut
 and Cuong C Nguyen who led me to the correct solution.

cuongc@nad.3com.com
drc@amherst.com
Boran.Sean@ch.swissbank.com
m1cxn01@FRB.GOV
daniel@sar3.CANR.Hydro.Qc.CA
Peter.Klein@um.erisoft.se
steveb@taheke.glas.apc.org
rgraham@cbetech.com.au
iss2jrs@uugtedc.gtedc.gte.com
mnoguchi%japan%csfb1@uunet.uu.net

The original query is below.
> To: sun-managers@eecs.nwu.edu
> Subject: Seagate Barracuda on Sparc 20 Sol 2.3
> Status: R
>
> Hello Managers,
>
> I have been trying to add a Seagate Barracuda drive to a Sparc 20. It all
> appears to go swimmingly until I mount a partition and try and use the disk.
> After about 5 minutes of use (I'm tarring user's accounts on to it) I get
> a series of messages as follows:
>
> Nov 29 19:33:05 kirk unix: WARNING: /iommu@f,e0000000/sbus@f,e0001000/espdma@f,400000/esp@f,800000 (esp0):Nov 29 19:33:05 kirk unix: No command for reconnect of Target 0 Lun 0
> Nov 29 19:33:08 kirk unix: WARNING: /iommu@f,e0000000/sbus@f,e0001000/espdma@f,400000/esp@f,800000/sd@0,0 (sd0):
> Nov 29 19:33:08 kirk unix: SCSI transport failed: reason 'reset':
> Nov 29 19:33:08 kirk unix: retrying command
> Nov 29 19:34:16 kirk unix: WARNING: /iommu@f,e0000000/sbus@f,e0001000/espdma@f,400000/esp@f,800000 (esp0):Nov 29 19:34:16 kirk unix: Disconnected tagged cmds (1) timeout for Targ
> Nov 29 19:34:19 kirk unix: et 0.0
> Nov 29 19:34:19 kirk unix: WARNING: /iommu@f,e0000000/sbus@f,e0001000/espdma@f,400000/esp@f,800000/sd@0,0 (sd0):
> Nov 29 19:34:19 kirk unix: SCSI transport failed: reason 'timeout'
> Nov 29 19:34:19 kirk unix: : retrying command
> Nov 29 19:34:19 kirk unix: WARNING: /iommu@f,e0000000/sbus@f,e0001000/espdma@f,400000/esp@f,800000/sd@0,0 (sd0):
> Nov 29 19:34:19 kirk unix: SCSI transport failed: reason 'reset':
> Nov 29 19:34:19 kirk unix: retrying command
> Nov 29 19:34:21 kirk unix: WARNING: /iommu@f,e0000000/sbus@f,e0001000/espdma@f,400000/esp@f,800000/sd@0,0 (sd0):
> Nov 29 19:34:21 kirk unix: Error for command 'write(10)' Error Lev
> Nov 29 19:34:21 kirk unix: el: Retryable
> Nov 29 19:34:21 kirk unix: Requested Block 254400, Error Block: 0
> Nov 29 19:34:21 kirk unix: Sense Key: Not Ready
> Nov 29 19:34:21 kirk unix: Vendor 'SEAGATE': ASC = 0x4 (drive not ready), ASCQ = 0x1, FRU = 0x1
> Nov 29 19:34:23 kirk unix: WARNING: /iommu@f,e0000000/sbus@f,e0001000/espdma@f,400000/esp@f,800000/sd@0,0 (sd0):
> Nov 29 19:34:23 kirk unix: Error for command 'write(10)' Error Lev
> Nov 29 19:34:23 kirk unix: el: Retryable
> Nov 29 19:34:23 kirk unix: Requested Block 254462, Error Block: 0
> Nov 29 19:34:23 kirk unix: Sense Key: Not Ready
>
> and the disk accesses stop. These messages also then appear for the other
> 3 disks in the chain. (2 internal 1GB, 1 external 2 GB and the barracuda).
>
> I have tried the disk on a sparc 5 with longer cables 3 other external devices
> and passive termination and it works perfectly. On the sparc 20 there is only
> 1 other external device and active termination.
>
> Any advice as to my next move would be gratefully received.
>
> Tom.
>
Tom.

-------------------------------------------------------------------------------
  Tom Crummey, EMAIL: tom@sos.bangor.ac.uk | /\
  University of Wales, Bangor, | / \/\
  School of Ocean Science, TEL: +44 (0)1248 382694 | /\/ \ \
  Menai Bridge, FAX: +44 (0)1248 716367 | / ======\=\
  Gwynedd, LL59 5EY, U.K. MOBILE: +44 (0) 585 379987 | B A N G O R
-------------------------------------------------------------------------------



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:09:17 CDT