a bug ?

Julian Anastasov ja at ssi.bg
Thu Oct 18 16:34:53 BST 2001


	Hello,

On Thu, 18 Oct 2001, Djamil ESSAISSI wrote:

> i checked dmesg for another reason ... i found this : too lucky the server still runs !
> i must have done a bad manipulation ... so i'm sending this just in case it's important.

	Yes, this is interesting.

> IPVS: Connection hash table configured (size=4096, memory=32Kbytes)
> IPVS: Each connection entry needs 120 bytes at least
> IPVS: ipvs loaded.
> IPVS: [wlc] scheduler registered.
> IPv4 over IPv4 tunneling driver
> IPVS: Scheduler module ip_vs_lwc.o not found
> Unable to handle kernel NULL pointer dereference at virtual address 00000010
>  printing eip:
> e08f7108
> *pde = 00000000
> Oops: 0000
> CPU:    0
> EIP:    0010:[<e08f7108>]    Not tainted
> EFLAGS: 00010296
> eax: 00000030   ebx: c0508ef0   ecx: c2f80000   edx: dea836e0
> esi: c0508ec0   edi: 00000000   ebp: fffffffe   esp: c2f81e30
> ds: 0018   es: 0018   ss: 0018
> Process ipvsadm (pid: 19493, stackpage=c2f81000)
> Stack: 00000016 00000006 c0508ec0 00001600 00000000 00000000 e08f7c90 c0508ec0
>        c2f81e70 e0900120 c2f81ed0 c0234794 bffffc50 00000000 e08ffb40 9ada2bd4
>        00000000 c01c2280 ca41aaa0 00000482 bffffc50 0000005c 00000002 ca41aaa0
> Call Trace: [<e08f7c90>] [<e0900120>] [<e08ffb40>] [<c01c2280>] [<c01c22da>]
>    [<c01cbfdf>] [<c01dfde8>] [<c01e5366>] [<c01b7e81>] [<c01b8526>] [<c0106b0b>]
>
> Code: 8b 47 10 85 c0 74 0a ff 48 10 8b 47 10 80 48 14 08 83 7c 24
>  ip_tables: (c)2000 Netfilter core team
> ip_conntrack (4086 buckets, 32688 max)

	Also build IPVS in the kernel, not as module. By this way
we have more chance to see symbol information after ksymoops. I
assume this is 0.9.4?

Regards

--
Julian Anastasov <ja at ssi.bg>





More information about the lvs-users mailing list