It'd be very interesting to know if this problem exists in -current.
Can you give it a try?

Also is there any visible pattern in corrupted data? Like last xxx bytes
of block, first xxx bytes of page?

	Andrey

On Wed, Nov 20, 2002 at 05:06:02PM -0700, Rick Kelly wrote:
> 
> I have a U5 with a 4gig Sun IDE disk and a SymBios SCSI controller with
> a Seagate 18gig Ultra SCSI disk.
> 
> I've been moving stuff around from IDE to SCSI by using the command:
> 
> sidewinder# cd dir1;tar cvf - . |(cd dir 2; tar xfBp -)
> 
> This seems to result in truncated or munged files at the end of the
> transfer.
> 
> Using cp seems to work all right.
> 
> I wonder if it is a pipe problem compounded by the difference in drive speed.
> 
> wd0 at pciide0 channel 0 drive 0: <ST34321A>
> wd0: drive supports 32-sector PIO transfers, LBA addressing
> wd0: 4103 MB, 8894 cyl, 15 head, 63 sec, 512 bytes/sect x 8404830 sectors
> wd0: drive supports PIO mode 4, DMA mode 2, Ultra-DMA mode 2 (Ultra/33)
> wd0(pciide0:0:0): using PIO mode 4, DMA mode 2 (using DMA data transfers)
> 
> sd0 at scsibus0 target 0 lun 0: <SEAGATE, ST318418N, 0003> SCSI3 0/direct fixed
> sd0: 19001 MB, 26434 cyl, 2 head, 736 sec, 512 bytes/sect x 38914049 sectors
> sd0: sync (50.0ns offset 16), 8-bit (20.000MB/s) transfers, tagged queueing
> 
> -- 
> Rick Kelly  rmk@rmkhome.com  www.rmkhome.com