On Mon, 12 Jul 2004, Martin Sarsale wrote: > Dear all: > > I have to copy 500GBs from an external lacie usb drive (NTFS) to my > HD. When I started I noticed that the thing was veeery slow (USB 1.x, > around 1mb/s) so I asked my ISP to add a USB 2.0 card on my server > (Dual Xeon 2.8, 1gb ram, Debian Stable running 2.4.26-686-smp). > > After they installed the new card I found that NTFS entered in some > infinite loops when doing a recursive copy from the USB drive to the > HD and I couldn't access all the files on the disk. > > When reading the logs, I saw that the errors where NTFS related (it > sounded rare to me, since the NTFS driver was the same, the only > difference was the USB card) so I had a conversation with linux-NTFS > people. After some tests, they pointed out that probably the problem > was USB related > (http://sourceforge.net/forum/forum.php?thread_id=1106787&forum_id=44085).
> Then, I installed kernel 2.6.7 (from kernel.org not debian's) to see > if things go better but it was the same as 2.6.6. > > People from linux NTFS thought it was a bug on their code so they > asked me to see if I could send some bytes from the disk to analyze > them (http://sourceforge.net/forum/message.php?msg_id=2660915) and > this is their answer > (http://sourceforge.net/forum/message.php?msg_id=2661050): > > " > Apparently you have a USB 2.0 problem. > > The data you sent are totally OK now. If it's compared with the wrong > one, the bytes start to differ at exactly position 256 (often used as > a unit during data transfers). Something messes up things randomly > before they would arrive to the NTFS driver. > " > > I would love to help you to fix this problem. I have full access > (remote) to the machine and it's not yet in production enviroment so > we can boot it the times we need. > > Any ideas? It's pretty unusual, although not unheard of, for data to be corrupted when read from a USB storage device. In this case you can get more information by turning on the usb-storage debugging option in your kernel's configuration. It generates a lot of output to the system log, so try running your test with something as small as you can manage while still triggering the bug -- in other words, don't try to transfer hundreds of GB with the debugging turned on! Perhaps one of those tests suggested by the NTFS people would be good. Something helpful might show up in your log. Alan Stern ------------------------------------------------------- This SF.Net email sponsored by Black Hat Briefings & Training. Attend Black Hat Briefings & Training, Las Vegas July 24-29 - digital self defense, top technical experts, no vendor pitches, unmatched networking opportunities. Visit www.blackhat.com _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-users
