Hi, On Thu, 2004-01-22 at 10:21, Mike McCormack wrote: > Yeah, it looks like I missed a bunch of messages in /var/log/syslog... > see the attached trace. > > The trouble appears to have started with this message: > Jan 22 01:52:42 kimchi kernel: EXT3-fs warning (device hda2): > ext3_unlink: Deleting nonexistent file (2874981), 0 The first error here is: > Jan 18 16:16:12 kimchi kernel: Unable to handle kernel paging request at virtual address 00200014 > Jan 18 16:16:12 kimchi kernel: EIP: 0060:[<c016ba37>] Not tainted > Jan 18 16:16:12 kimchi kernel: EIP is at prune_dcache+0x177/0x200 > Jan 18 16:16:12 kimchi kernel: eax: 00200000 ebx: ca2c8cc0 ecx: c266e364 edx: dfe3a000 > Jan 18 16:16:12 kimchi kernel: esi: c266e354 edi: 00000018 ebp: dfe3be7c esp: dfe3be54 which on first glance looks like memory corruption: there's a single bit-flip in %eax and we oops trying to dereference that. Then on reboot we detect an inconsistency on disk, and ext3 behaves as expected: it aborts the journal, remounts readonly, and you see the filesystem marked for fsck on the next reboot. Have you seen this more than once? This footprint usually indicates a hardware problem, and memtest86 is a useful tool to diagnose that. --Stephen _______________________________________________ Ext3-users@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/ext3-users
