Am Donnerstag, 30. Mai 2013, 23:10:23 schrieb Michal Kubecek:
On Thu, May 30, 2013 at 10:38:10PM +0200, Günther J. Niederwimmer wrote:
Am Donnerstag, 30. Mai 2013, 16:00:58 schrieb Michal Kubeček:
On Thursday 30 of May 2013 15:43EN, Jiri Slaby wrote:
On 05/30/2013 10:57 AM, Günther J. Niederwimmer wrote:
techz kernel: [ 2632.076419] IPv4: martian source 192.168.100.255 from 192.168.100.100, on dev eth1 techz kernel: [ 2632.076425] ll header: 00000000: ff ff ff ff ff ff 70 71 bc e2 cf fb 08 00 ......pq......
2: eth0:
mtu 1500 qdisc pfifo_fast state UP qlen 1000 link/ether 70:71:bc:e2:cf:fb brd ff:ff:ff:ff:ff:ff inet 192.168.100.100/24 brd 192.168.100.255 scope global eth0 inet6 2001:15c0:65ff:868c:bd3d:4977:e126:5f91/64 scope global temporary
dynamic
valid_lft 585359sec preferred_lft 66359sec
inet6 2001:15c0:65ff:868c:7271:bcff:fee2:cffb/64 scope global dynamic
valid_lft 2591515sec preferred_lft 604315sec
inet6 fe80::7271:bcff:fee2:cffb/64 scope link
valid_lft forever preferred_lft forever
3: eth1:
mtu 1480 qdisc pfifo_fast state UP qlen 1000 link/ether 70:71:bc:e2:cf:fd brd ff:ff:ff:ff:ff:ff inet6 2001:15c0:65ff:868c:a46a:6dcc:3c76:cb42/64 scope global temporary
dynamic
valid_lft 585359sec preferred_lft 66359sec
inet6 2001:15c0:65ff:868c:7271:bcff:fee2:cffd/64 scope global dynamic
valid_lft 2591515sec preferred_lft 604315sec
inet6 2001:15c0:65ff:868c::64/64 scope global
valid_lft forever preferred_lft forever
inet6 fe80::7271:bcff:fee2:cffd/64 scope link
valid_lft forever preferred_lft forever
So the problem is that a (broadcast) packet from you sent to eth0 was received on eth1 which RP filter doesn't like. I dare to guess that you connected both eth0 and eth1 to the same ethernet segment. It is is not necessarily wrong and there are legitimate reasons to do that (but perhaps bonding might be more suitable solution). Anyway, in such case (as in many others), RP filter should be either disabled or at least set to loose mode, i.e. rp_filter sysctl should be set to 0 or 2. Unfortunately the default in OpenSuSE is 1 (strict mode).
OK, thak's I search in google ;) This is for me, a mystery Situation. I make yesterday a Test, I installed the system new only with one NIC (no seperate IPv6), and now the "martian_source" Log is gone again (?). Have any a link for Information or tests for this situation, I have a XEN SLES Test Installation with the same Problem, full running logs on all Instances :-(. I don't remember is this starting with a kernel update or my IPv6 config? -- mit freundlichen Grüßen / best Regards, Günther J. Niederwimmer -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org