Public bug reported: Hello. I am reporting this for e2fsprogs, even if this may in fact be a more general problem, perhaps only solvable in the kernel - here are the symptoms:
I am facing a situation with most likely faulty hardware, either the HDD or the main board, and the system seems to be unable to cope with it happily. Running "e2fsck -c" most clearly demonstrates the problem. The I/O error(s) first occurred when trying to read a recently-written file from the HDD - AmaroK just choked and crashed on the file, "cp" and "dd" of that file failed in the middle (constantly with the same amount of bytes retrieved). With "cp" and "dd", the system would report an I/O error when failing. I then tried unmounting the partition and running "e2fsck -c" on it. It is a 5 GB partition - the check ran fine for the first 4 GB, but thereafter apparently hit the area that can't be reached, and started giving I/O errors, reporting a new failed sector each time, whose number was 8 larger than the previous. At this point, the system became terribly unresponsive / hung, and e2fsck couldn't be Ctrl-C'ed. I tried to repeat "e2fsck -c" by stopping X and doing a console login. The same symptoms, with the messages about I/O errors filling up each virtual terminal that I had logged into. (I was unable to kill e2fsck from another VT, but I think it _might_ have been possible when the problems started, had I had the command ready on the other terminal. I do not know what can be done at the face of faulty hardware. I do hope, though, that a better behaviour is possible. ** Affects: e2fsprogs (Ubuntu) Importance: Undecided Status: Unconfirmed -- e2fsck / linux kernel chokes on I/O errors https://launchpad.net/bugs/64914 -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs