I forgot to tell that I've seen this problem with kernels as far back as 2.6.8. My feeling is that it happens more frequently now than half a year ago. I know this points in the direction of hardware problems, but Windows works, even for large file copies. The vendor mounted two disks and copied a 15 GB dummy file back and forth. Result OK. Test was repeated while system was stressed. Still OK. Finally a continuous 65 hour stress test.
May be marginal timings have become more problematic over time. I know this will be hard to track down. Suggestions for workarounds are welcome. Is there anything like hdparm for SATA? Regards Jon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]