One of our servers (false) just oopsed. In the middle of lunch. Any advise? On console: false kernel: Assertion failure in journal_start() at transaction.c line 245: "handle->h_transaction->t_journal == journal" In kernel log (ran it through ksymoops): Nov 13 12:35:37 false kernel: Unable to handle kernel NULL pointer dereference at virtual address 00000000 Nov 13 12:35:37 false kernel: current->tss.cr3 = 00a2b000, %%cr3 = 00a2b000 Nov 13 12:35:37 false kernel: *pde = 00000000 Nov 13 12:35:37 false kernel: Oops: 0002 Nov 13 12:35:37 false kernel: CPU: 1 Nov 13 12:35:37 false kernel: EIP: 0010:[journal_start+87/232] Nov 13 12:35:37 false kernel: EFLAGS: 00010286 Nov 13 12:35:37 false kernel: eax: 00000071 ebx: c5d25da0 ecx: 00000246 edx: cfff8000 Nov 13 12:35:37 false kernel: esi: c4b82000 edi: cffe3b60 ebp: 0000006c esp: c4b83df0 Nov 13 12:35:37 false kernel: ds: 0018 es: 0018 ss: 0018 Nov 13 12:35:37 false kernel: Process mk-slides (pid: 19934, process nr: 161, stackpage=c4b83000) Nov 13 12:35:37 false kernel: Stack: c023bce3 000000f5 c023be3f c027a758 c9559188 00000000 00000000 c014aa2e Nov 13 12:35:37 false kernel: cffe3b60 0000006c c027a758 c9559188 c4b83e9c c010fdfc 00000246 cc66dc00 Nov 13 12:35:37 false kernel: c014a70a c9559188 c013790b c9559188 cc66dc00 c9559188 c2a70420 c0135e42 Nov 13 12:35:37 false kernel: Call Trace: [cprt+18694/54599] [cprt+19042/54599] [ext3_delete_inode+62/392] [smp_apic_timer_interrupt+16/24] [ext3_put_inode+10/16] [iput+155/576] [dput+166/328] Nov 13 12:35:37 false kernel: Code: c6 05 00 00 00 00 00 83 c4 14 ff 43 08 89 d8 5b 5e 5f 5d c3 Code; 00000000 Before first symbol 00000000 <_EIP>: Code; 00000000 Before first symbol 0: c6 05 00 00 00 00 00 movb $0x0,0x0 Code; 00000007 Before first symbol 7: 83 c4 14 add $0x14,%esp Code; 0000000a Before first symbol a: ff 43 08 incl 0x8(%ebx) Code; 0000000d Before first symbol d: 89 d8 mov %ebx,%eax Code; 0000000f Before first symbol f: 5b pop %ebx Code; 00000010 Before first symbol 10: 5e pop %esi Code; 00000011 Before first symbol 11: 5f pop %edi Code; 00000012 Before first symbol 12: 5d pop %ebp Code; 00000013 Before first symbol 13: c3 ret Additional information: <perbu@false:~> uname -a Linux false.linpro.no 2.2.20-ext3-0.0.7a #1 SMP Fri Nov 9 17:03:49 CET 2001 i686 unknown -- Per Andreas Buer
Hi, On Tue, Nov 13, 2001 at 01:13:57PM +0100, Per Andreas Buer wrote:> false kernel: Assertion failure in journal_start() at transaction.c line > 245: "handle->h_transaction->t_journal == journal"Looks like a recursive transaction leaked. That would normally be an indication of some odd recursion in the VM, but:> In kernel log (ran it through ksymoops): > > Nov 13 12:35:37 false kernel: Unable to handle kernel NULL pointer dereference at virtual address 00000000 > Nov 13 12:35:37 false kernel: Call Trace: [cprt+18694/54599] [cprt+19042/54599] [ext3_delete_inode+62/392] [smp_apic_timer_interrupt+16/24] [ext3_put_inode+10/16] [iput+155/576] [dput+166/328]... with only this one line of call trace, I can't see where the fs is being called. I would normally expect the call trace to show where ext3 was being called from (not just dput, that must be being called from somewhere else). Without that context, it's hard to see where we might have gone wrong. Cheers, Stephen