Initial Post:
> From ra.mcs.anl.gov!sun-managers-relay Wed Nov 16 23:11:01 1994
> Sender: sun-managers-relay@ra.mcs.anl.gov
> Message-Id: <m0r7mC1-0008BiC@hobbes.crc.com>
> Date: Wed, 16 Nov 94 10:22 EST
> I am frequently receiving the following error:
>
> Nov 12 05:12:27 fred unix: polled command timeout
> Nov 12 05:12:27 fred unix: esp: State=CLEARING Last State=CLEARING
> Nov 12 05:12:27 fred unix: esp: Latched stat=0x97<IPND,XZERO,MSG,CD,IO>
> intr=0x8<FCMP> fifo 0x80
> Nov 12 05:12:27 fred unix: esp: last msg out: NO-OP; last msg in: COMMAN
> D COMPLETE
> Nov 12 05:12:27 fred unix: esp: DMA csr=0x40040010<INTEN>
> Nov 12 05:12:27 fred unix: esp: addr=fc008bea dmacnt=0 last=fc008be8 las
> t_cnt=14
> Nov 12 05:12:27 fred unix: esp: Cmd dump for Target 6 Lun 0:
> Nov 12 05:12:27 fred unix: esp: cdblen=6, cdb=[ 0x3 0x0 0x0 0x0 0x14 0x0
> ]; Status=0x0
> Nov 12 05:12:27 fred unix: esp: pkt_state=0x1b<STS,XFER,SEL,ARB> pkt_fla
> gs=0x8403 pkt_statistics=0x0
> Nov 12 05:12:27 fred unix: esp: cmd_flags=0x1422 cmd_timeout=60
> Nov 12 05:24:33 fred unix: polled command timeout
>
> Sparc Center 2000
> Solaris 2.3
> 2.3Recommended_Patches
I haven't had a chance to check any of this because the machines in question
are off-site. I will be checking them later. The problem seems to have
gone away for now. I was getting this error on 3 new machines.
Here are the responses (Thanks for the help):
From: fm@Germany.Sun.COM (Florian Meyer - Sun Germany Product Support - Munich)
Bug Id: 1172254
Category: kernel
Subcategory: driver
Release summary: 4.2
Synopsis: 'polled command timeout' error; can't determine device state
Integrated in releases:
Patch id:
Description:
A 'polled command timeout' on a disk drive during a system boot needs additional
information displayed so that the Vendor, Product and Revision Level of the devicecan be determined from the log data.
(This is the info from SunSolve.)
From: "M.Masarati Cadds/Unix Support +39 2 26913370)" <M.MASARATI@ITMILB.cv.comI've just seen this message and the problem was the ethernet cable :
the yellow cable of the network has 2 terminators at the begin and at the
end of the cable,If you got a tester try to test the electric resistance
of the terminators, it must be 50 ohm.
Also try on the tranceiver, the resistance must be 25 ohm .
I've got this problem because my tranceiver resistance wasn't 25 ohm,so it
caused network problems on the net with this console message.
From: Curt Freeland <curt@ecn.purdue.edu>
Install 101378-09. This should fix the problem (at least it did for us).
From: "Srinivasa R. Yalavarthy" <srini@concorde.com>
this is definitely due to bad scsi. You need to check your scsi cable first, then check
for scsi terminator. You should have one if you have a daisy chain on your scsi. It seems like
target 6 is the culprit, which is usually cdrom. So I would check all those things then if the
problem still persists, I would check scsi fuse on the mother board.
From: bismark@alta.jpl.nasa.gov (Bismark Espinoza)
You may have SCSI-2 and "fast SCSI-2" devices in some
of your SCSI chains. There may be communication
problems from scsi controller to/from device.
A previous letter referred to patch #101378-09 that
supposedly fixes it. Another fix is to have chains
with only one type of peripheral.
From: Richard_Ravich@internet.microp.com (Richard Ravich)
It appears to me that what is causing the problem should be reported before the
error messages that you sent out. I'd suggest that there may be a problem with
the SCSI code on the drive (I hope it is not one of ours!) or a possible problem
in termination or cabling.
For what it's worth, the command that is giving the problem is a request sense
command, and as far as I can tell, should work.
From: "David M. Di Gioia" <David.Digioia@gain.com>
Perhaps they are FAST-SCSI-related errors.
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:09:16 CDT