Fuji 2263 "reset" problems
2007-12-25 7:21:00
11:04:23 sparky vmunix: esp0: current command timeout for Target 0 Lun 0
11:04:23 sparky vmunix: sd3: SCSI transport failed: reason 'reset': retrying
command
11:04:23 sparky vmunix: sd0: SCSI transport failed: reason 'reset': retrying
command
The solution as suggested by several people was to disable the
"read-ahead cache" on the disk.
One of the clearest discriptions came from "sharp@naoa.edu" as follows.
------------begin included---------------------
If you have the technical handbook, turn to page 3-8 and
turn off the read-ahead caching. If you don't, then ...
At the back of the board, below the SCSI connectors, there's a
"terminal" referred to as CN9. Low numbers are at the bottom,
furthest from the connector
SCSI connectors |
13 __
--| | | 13 11 9 7 5 3 1
--| | | | | | | | |
--| | | 14 12 10 8 6 4 2
--| | ------circuit board------
--| | | ^
--| | |
--|__| | 5&6 - read-ahead
1
As shipped, pins 5 and 6 are shorted, enabling the feature.
Open this connection and all should be well.
------------end included---------------------
Other suggestions included short cable lengths, running format, swapping
drives, and adding terminators to each drive.
Thanks to
"D.W. Desousa" <dan@esker.geog.utoronto.ca>
"Hilarie K. Orman" <ho@la.tis.com>
Fabrice Le Metayer <fabrice@sj.ate.slb.com>
Jim Guyton <guyton%condor@rand.org>
Stefan Mochnacki <stefan@centaur.astro.utoronto.ca>
Steve Hanson <hanson@pogo.fnal.gov>
Steve.Piette@Central.Sun.COM (Steve Piette Midwest Area SE)
andrew@calvin.doc.ca (Andrew Patrick)
censun1!fox!cb@uunet.UU.NET (Charlie Butterfield)
frl@phebos.aps.anl.gov (Frank Lenkszus)
kindred@rooster.acn.purdue.edu (Don Kindred)
mike@sparky.uwaterloo.ca (Michael Scott)
mis@seiden.com (Mark Seiden)
phil@dgbt.doc.ca (Phil Blanchfield)
root@rainbow.cse.nau.edu (System Administrator)
sgf@cfm.brown.edu (Sam Fulcomer)
sharp@noao.edu (Nigel Sharp)
and any others I may have missed.
--Gene
Comments
Got something to say?
You must be logged in to post a comment.
