[crash] 2.0.32 nfs_fhget crashed webserver

Tore Olsen (toreo@tihlde.hist.no)
02 Mar 1998 14:25:35 +0100

Our webserver went crazy today, scrolling messages on the screen which
I couldn't read. The following was logged to syslog twice before I
hard reset the box:
stack segment: 0000
CPU: 0
EIP: 0010:[nfs_fhget+102/212]
EFLAGS: 00010286
eax: 001edf18 ebx: 00000000 ecx: 052d9d00 edx: 0d3af97b
esi: 001edf18 edi: 052d9d00 ebp: ffffff9c esp: 01d12e38
ds: 0018 es: 0018 fs: 002b gs: 002b ss: 0018
Process httpd (pid: 8559, process nr: 29, stackpage=01d12000)
Stack: 00000000 01d12e9c 052d9d00 01d12f28 01d12f28 01d12f08 00000000 00165f57
00000000 01d12e9c 052d9d00 01d12ec4 01d10008 00169ccd 052d9d00 01d12e9c
00000000 01d12e9c 052d9d00 01d12f28 00169fa1 00169fb8 001edf18 01d12f08
Call Trace: [nfs_proc_lookup+439/452] [nfs_lookup_cache_add+73/264] [nfs_lookup+325/408] [nfs_lookup+348/408] [lookup+222/248] [_namei+91/228] [namei+48/72]
[sys_newstat+41/88] [system_call+85/124]
Code: 8b 45 28 6a 01 50 56 e8 c6 b3 fb ff 89 c3 83 c4 0c 85 db 75

Running ksymoops gave this, which I'm not sure is very helpful, but included
Code: 00000000 <_EIP> movl 0x28(%ebp),%eax
Code: 00000003 <_EIP+3> pushl $0x1
Code: 00000005 <_EIP+5> pushl %eax
Code: 00000006 <_EIP+6> pushl %esi
Code: 00000007 <_EIP+7> call fffbb3d2 <_EIP+fffbb3d2>
Code: 0000000c <_EIP+c> movl %eax,%ebx
Code: 0000000e <_EIP+e> addl $0xc,%esp
Code: 00000011 <_EIP+11> testl %ebx,%ebx
Code: 00000013 <_EIP+13> jne 00000015 <_EIP+15>
Code: 00000015 <_EIP+15> nop
Code: 00000016 <_EIP+16> nop
Code: 00000017 <_EIP+17> nop

This has never happened before. /home is mounted via NFS, so I guess
this happened when some www-robot tried to index lots of /~* URLs.



- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu