http://bugzilla.novell.com/show_bug.cgi?id=600813 http://bugzilla.novell.com/show_bug.cgi?id=600813#c0 Summary: External USB disk detects IO errors while flushing file data Classification: openSUSE Product: openSUSE 11.2 Version: Final Platform: i386 OS/Version: openSUSE 11.2 Status: NEW Severity: Normal Priority: P5 - None Component: Basesystem AssignedTo: bnc-team-screening@forge.provo.novell.com ReportedBy: joerg.dippel@gmx.de QAContact: qa@suse.de Found By: --- Blocker: --- I've seen a problem copying large files between two USB disks with an ext3 file system on SuSE 10.0. The problem appears if a crypted file system of size around 500 GB is copied by root to a non crypted ext3 file system. A similar problem has been reproduced on SuSE 11.2 x86_64 for ext3. The file system is created with mkfs (util-linux-2.12q-26 - SuSE 10.0). The case is rather simple. Some directories are copied recursively to the other device. Suddenly the target device is umounted and remounted with another naming while the copying is still in progress. There are already two related bugs reported (bug105585 and especially bug113827). At the moment I cannot exclude a technical/hardware defect for the target USB disk, but different OSs do not detect any errors on hardware. Already I've used a disk test program by Heise. The used kernel is 2.6.13-15.18-default on an IBM ThinkPad T41p. bug105585,comment19 names a previous kernel as a resolution for the bug. Regression is done on an Acer Aspire 8930G: Linux linux-cice 2.6.31.12-0.2-desktop #1 SMP PREEMPT 2010-03-16 21:25:39 +0100 x86_64 x86_64 x86_64 GNU/Linux A trial to reproduce the case with another but similar test setup can reproduce the umount and remount. cp and diff do not return anything to standard error but syslogd reports: Message from syslogd@linux-cice at Apr 29 10:01:00 ... kernel:[ 3889.880861] journal commit I/O error The new test setup is two partitions on the target USB disk. First partition hold the data with no error visible. The data is copied to the second partition, but after 54% filling the observation occurs: /dev/sdd2 730572624 373383128 320078544 54% The test case is now, only to remove the failed file and to recopy it from first to second partition. The cleared output of dmesg follows. If it's a problem on handling bad blocks it's okay. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.