[EMAIL PROTECTED]:~# uname -a Linux om 2.6.22-14-generic #1 SMP Sun Oct 14 23:05:12 GMT 2007 i686 GNU/Linux [EMAIL PROTECTED]:~# cat /proc/version_signature Ubuntu 2.6.22-14.46-generic
I was copying disks with rsync. I had two 1-TB SATA drives involved, one installed inside the chassis, the other on an eSATA cable to an external box. Both of these drives performed fine - no bugs. Both are formatted with ext2 filesystems. I have borrowed two USB/firewire external drives, 1 TB each, formatted with NTFS filesystems. The drives are made by "Eagle Consus" model number ET-CSDU2F-BK. Each of these drive boxes contains a small SATA RAID controller set up in "JBOD" mode, holding two 500GB drives. I was using "rsync -rav --inplace /mnt/jon1/ /gd1/" to copy from a USB disk to a SATA disk. Rsync would copy along at about 20 MBytes/sec (according to the iostat command). It chugged along fine for minutes or hours. I discovered that I was able to get about 30 MBytes/sec if I ran rsync simultaneously from the second USB disk to the second SATA disk. So I started doing that. But I got several inexplicable hangs. rsync would be copying along, and would suddenly stop. That was the symptom; no error message, no crash, just a hang on one rsync, while the other would go on. When it would hang, it was always /mnt/jon1 that would hang. The rsync command runs as two or three processes; one of those would hang in "D+" state forever: 9947 pts/3 D+ 28:36 | \_ rsync -rav --inplace /mnt/jon1/ /gd1/ "ps -eopid,tt,user,fname,tmout,f,wchan" shows it waiting in sync_page: 9947 pts/3 gnu rsync - 0 sync_page When I do "ls -l /proc/9947/fd", I get a hang. Luckily I can break out of it by "kill -9" from another terminal. So I snuck up on it: [EMAIL PROTECTED]:~# ls -l /proc/9947/fd Killed [EMAIL PROTECTED]:~# /bin/ls /proc/9947/fd 0 1 2 3 4 5 [EMAIL PROTECTED]:~# /bin/ls -l /proc/9947/fd/0 lrwx------ 1 gnu uucp 64 2007-11-14 23:28 /proc/9947/fd/0 -> /dev/pts/3 [EMAIL PROTECTED]:~# /bin/ls -l /proc/9947/fd/1 lrwx------ 1 gnu uucp 64 2007-11-14 23:28 /proc/9947/fd/1 -> /dev/pts/3 [EMAIL PROTECTED]:~# /bin/ls -l /proc/9947/fd/2 lrwx------ 1 gnu uucp 64 2007-11-11 18:25 /proc/9947/fd/2 -> /dev/pts/3 [EMAIL PROTECTED]:~# /bin/ls -l /proc/9947/fd/3 lr-x------ 1 gnu uucp 64 2007-11-14 23:28 /proc/9947/fd/3 -> /mnt/jon1/The_Grateful_Dead/1973-11-14/02-Sugaree.flac (deleted) [EMAIL PROTECTED]:~# /bin/ls -l /proc/9947/fd/4 lrwx------ 1 gnu uucp 64 2007-11-14 23:28 /proc/9947/fd/4 -> socket:[70521] I was able to do "lsusb" without a hang, but "lsusb -v" hangs unkillably. Unplugging the USB device doesn't resolve the problem. In fact, after resetting power on the computer to escape, the BIOS would hang if this drive was plugged into USB. When I pulled it out and reinserted it, the boot went on, but after booting, I got errors: "usb 2-9: device descriptor read/64, errno -110." and "usb 2-9: device descriptor read/8, errno -110." It required power-cycling the external disk box to reset it from this state. I can continue to borrow these external drives for another week or two, and the problem seems pretty reproducible. I don't mind if you want me to try running a test kernel, or to do some other actions to help run down this bug. Computers ought to work reliably, even under heavy loads. ** Attachment added: "dmesg output" http://launchpadlibrarian.net/10375119/hang.dmesg -- One broken USB storage device can hang the entire USB subsystem https://bugs.launchpad.net/bugs/136822 You received this bug notification because you are a member of Ubuntu Bugs, which is the bug contact for Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
