Description of Problem: Got the following error from EXT3 using the 2.4.18-4 kernel on a dual CPU box with 4 disks (arranged as a software RAID mirrored pairs). Jun 10 10:30:41 purple kernel: Assertion failure in journal_write_metadata_buffer() at journal.c:406: "buffer_jdirty(jh2bh(jh_in))" Jun 10 10:30:41 purple kernel: ------------[ cut here ]------------ Jun 10 10:30:41 purple kernel: kernel BUG at journal.c:406! Jun 10 10:30:41 purple kernel: invalid operand: 0000 Jun 10 10:30:41 purple kernel: autofs eepro100 ipchains ext3 jbd raid1 aic7xxx sd_mod scsi_mod Jun 10 10:30:41 purple kernel: CPU: 0 Jun 10 10:30:41 purple kernel: EIP: 0010:[<f885d9b4>] Not tainted Jun 10 10:30:41 purple kernel: EFLAGS: 00010282 Jun 10 10:30:41 purple kernel: Jun 10 10:30:41 purple kernel: EIP is at journal_write_metadata_buffer [jbd] 0x74 (2.4.18-4smp) Jun 10 10:30:41 purple kernel: eax: 0000001d ebx: 00000000 ecx: c02eeec0 edx: 000aa050 Jun 10 10:30:41 purple kernel: esi: 00000000 edi: ee129a40 ebp: f7dc8490 esp: f79afe44 Jun 10 10:30:41 purple kernel: ds: 0018 es: 0018 ss: 0018 Jun 10 10:30:41 purple kernel: Process kjournald (pid: 141, stackpage=f79af000) Jun 10 10:30:41 purple kernel: Stack: f8862181 00000196 000016dc f772a600 000000 00 00000000 f7dc8b20 00000000 Jun 10 10:30:41 purple kernel: ee129a40 f7dc8490 f885adf4 ee129a40 f7dc8b 20 f79afe98 000018e8 00000000 Jun 10 10:30:41 purple kernel: 00000001 00000ff4 d3e9e00c 00000001 ee129a 40 f6afd8b0 000018e8 00000001 Jun 10 10:30:41 purple kernel: Call Trace: [<f8862181>] .rodata.str1.1 [jbd] 0x4 e1 Jun 10 10:30:42 purple kernel: [<f885adf4>] journal_commit_transaction [jbd] 0x7 e4 Jun 10 10:30:42 purple kernel: [<c0119048>] schedule [kernel] 0x348 Jun 10 10:30:42 purple kernel: [<f885d806>] kjournald [jbd] 0x136 Jun 10 10:30:42 purple kernel: [<f885d6b0>] commit_timeout [jbd] 0x0 Jun 10 10:30:42 purple kernel: [<c0107286>] kernel_thread [kernel] 0x26 Jun 10 10:30:42 purple kernel: [<f885d6d0>] kjournald [jbd] 0x0 Jun 10 10:30:42 purple kernel: Jun 10 10:30:42 purple kernel: Jun 10 10:30:42 purple kernel: Code: 0f 0b 5e 5f 8b 7c 24 28 8b 4f 0c 85 c9 74 2 e c7 44 24 0c 01 The machine in question was still accepting input at the command line, however shutdown command failed (because of broken filesystem?) and required powercycle. No errors reported by fsck -f on any filesystem. Version-Release number of selected component (if applicable): kernel-smp-2.4.18-4 How Reproducible: Not very I'm afraid. We have three identical machines all running Redhat 7.3 for a number of weeks without any problems. The machine in question has been fine since it was rebooted on Monday. Just thought that I ought to report a potential problem with EXT3. I found the following on the ext3 developers list. Its the same assertion failure but a different kernel version: I don't know if this is relevant. https://listman.redhat.com/pipermail/ext3-users/2002-May/003587.html
Thanks. There has been one other possible report of this on 2.4.18-4smp, but I don't have it tracked down yet. If this becomes reproducible, would you be willing to try a debugging kernel to help catch the problem?
Yes, certainly.
I have found a possible cause for this problem, and it will be in a forthcoming errata. If you want, I can give you a copy of the patch to try it out; but if you are not able to reproduce the problem, it's probably not worth it.
There is a patched kernel on http://people.redhat.com/arjanv/testkernels/ if you want to try it --- it's 2.4.18-7.
*** This bug has been marked as a duplicate of 68026 ***