From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4a) Gecko/20030401 Description of problem: We are currently writing about 1MB per second of data, and creating and deleting about 600+ inodes per second. The following OOPS shows that the problem occurs in the 2.4.20-18.7smp kernel. Tracing the code, it appears that if inode_has_buffers(inode) returns true then the BUG is thrown. We are running with ext3 fs, and my guess would be that the problem exists in one of the following situations: The process grows, to the point where it is using more memory, and it tickles something in the virtual memory which corrupts the inode->i_dirty_buffers list so that the i_dirty_buffers are in fact not empty. ( This is a conjecture based on the fact that the ext3 journaling doesn't store the i_dirty_buffers info with the inode itself ). The other possibility is that such a high I/O rate is somehow corrupting the journaling, causeing the same type of issue. ------------[ cut here ]------------ kernel BUG at inode.c:126! invalid operand: 0000 e1000 ipt_REJECT ipt_state ip_conntrack iptable_filter ip_tables usb-ohci usbcore ext3 jbd aacraid sd_mod scsi_mod CPU: 0 EIP: 0010:[<c015b840>] Not tainted EFLAGS: 00010202 EIP is at destroy_inode [kernel] 0x10 (2.4.20-18.7smp) eax: 00000001 ebx: d913ab80 ecx: 00000001 edx: d913ab80 esi: d913ab80 edi: f8859b00 ebp: dd330a80 esp: c4667ef8 ds: 0018 es: 0018 ss: 0018 Process ecelerity (pid: 17532, stackpage=c4667000) Stack: d913ab80 c015d0f5 d913ab80 ffffffff e41d2980 dd330b00 c024a7cc dd330a80 d913ab80 e41d2980 c015aec6 d913ab80 dd330b00 c4666000 fffffff0 e41d2980 c4666000 00000000 c01530f0 dd330a80 00000000 d6706780 c4667f9c e41d2980 Call Trace: [<c015d0f5>] iput [kernel] 0x275 (0xc4667efc)) [<c015aec6>] d_delete [kernel] 0x66 (0xc4667f20)) [<c01530f0>] vfs_unlink [kernel] 0x1e0 (0xc4667f40)) [<c0150b80>] cached_lookup [kernel] 0x10 (0xc4667f58)) [<c0151c3a>] lookup_hash [kernel] 0x4a (0xc4667f68)) [<c01531b9>] sys_unlink [kernel] 0x89 (0xc4667f88)) [<c0108be3>] system_call [kernel] 0x33 (0xc4667fc0)) Code: 0f 0b 7e 00 ba 95 25 c0 8b 83 9c 00 00 00 8b 40 20 8b 40 04 Version-Release number of selected component (if applicable): 2.4.20-18.7smp How reproducible: Couldn't Reproduce Steps to Reproduce: Not really reproducible, has happened twice though. Actual Results: Same thing, kernel oops. Expected Results: No kernel oops. Additional info:
Thanks for the bug report. However, Red Hat no longer maintains this version of the product. Please upgrade to the latest version and open a new bug if the problem persists. The Fedora Legacy project (http://fedoralegacy.org/) maintains some older releases, and if you believe this bug is interesting to them, please report the problem in the bug tracker at: http://bugzilla.fedora.us/