On Sat, Dec 14, 2002 at 02:34:45PM +1000, Ray Phillips wrote:
> I'm running NetBSD 1.6 on a Pentium II 350 MHz machine with two IDE 
> disks (one's a master on the primary controller, the other's a master 
> on the secondary), both of which are practically new.  The system had 
> been behaving nicely and I was almost ready to put it into production 
> use.
> 
> However, I thought I'd make use of three old SCSI disks so I used ccd 
> to concatenate them with no striping.  (My intention was to connect 
> them all to a single SCSI controller when I got a ribbon cable with 
> four connectors, but for the time being I connected one disk to one 
> controller and two to another.)  I made sure to offset the ccd 
> partition on each disk by one cylinder, and made their filesystem 
> type ccd.  The contents of /etc/ccd.conf were:
> 
> # ccd   ileave  flags   component devices
> ccd0    0       none    /dev/sd0e /dev/sd1e /dev/sd2e
> 
> Executing 'ccdconfig -vC' seemed to work, as did  'newfs 
> /dev/rccd0e'.  I mounted the partition with 'mount /dev/ccd0e /mnt', 
> copied a directory (named photos, about 550 MB in size) to /mnt and 
> ran 'diff -r photos /mnt/photos' which said one of the binary files 
> was different but it didn't complete its tests because the system 
> froze and I had to power cycle it.  These messages were on the 
> console before I did:
> 
> sd1(trm0:0:1:0): SCSI OpCode 0x0a timed out
> trm0: over/under run error
> sd1(trm0:0:1:0): generic HBA error
> ccd0: error 5 on component 1
> 
> Now when I power up the machine (and during use) errors like this 
> appear on the console:
> 
> wd0: transfer error, downgrading to Ultra-DMA mode 1
> wd0(pciide0:0:0): using PIO mode 4, Ultra-DMA mode 1 (using DMA data 
> transfers)
> wd0a: error reading fsbn 1051232 of 1051232-1051233 (wd0 bn 1051295; 
> cn 1042 tn 15 sn 14), retrying
> wd0: (uncorrectable data error)
> wd0: transfer error, downgrading to DMA mode 2
> wd0(pciide0:0:0): using PIO mode 4, DMA mode 2 (using DMA data transfers)
> wd0a: error reading fsbn 1051232 of 1051232-1051233 (wd0 bn 1051295; 
> cn 1042 tn 15 sn 14), retrying
> wd0e: error reading fsbn 41177552 of 41177552-41177567 (wd0 bn 
> 46465520; cn 46096 tn 11 sn 59), retrying
> wd0: (uncorrectable data error)
> wd0: soft error (corrected)
> 
> Do you think the fact that these errors started immediately I began 
> to use a ccd partition was just a bizarre coincidence or are they 
> linked somehow?  The impression I got from scanning the mail archives 
> is ccd is mature and robust these days.

Are your SCSI disks internal or external ?
If internal it's possible that your power supply it too short.

-- 
Manuel Bouyer <bouyer@antioche.eu.org>
     NetBSD: 23 ans d'experience feront toujours la difference
--