I have a problem with one box which I just upgraded to 7.2 (it worked perfectly fine with 7.1 just before). Now I get messages like --------------------------------------- Nov 8 05:10:06 ws3 kernel: attempt to access beyond end of device Nov 8 05:10:06 ws3 kernel: 03:01: rw=0, want=1350676760, limit=3068383 Nov 8 05:10:06 ws3 kernel: EXT3-fs error (device ide0(3,1)): ext3_readdir: directory #371873 contains a hole at offset 32768 ---------------------------------------- over and over in the logs and then the kernel will eventually crash.
The error you reported here is a sign that something is corrupt, but it can mean anything: bad memory, a bad disk, disk controller dying, CPU cache errors, or random memory scribbles from the kernel. In the past it has most often been associated with hardware errors on the motherboard north or south bridge chipsets. What is your hardware config (especially CPU, motherboard, IDE controller and disk type)? Which kernel are you using? When you say the kernel will "eventually crash", what do you mean by that?
Some hardwareinfor: cpu: AMD Athlon(tm) Processor 700MHz IDE interface: VIA Technologies, Inc. Bus Master IDE (rev 10) HD: ST310212A kernel-2.4.9-13 The system was dead and would only react to the reset button. I looked into the log and the above massages were the last thing in there. However, it did not happen again since then. I'll close this bug since it doesn't seem reproducible right now.