From mboxrd@z Thu Jan 1 00:00:00 1970 From: Marcin Giedz Subject: Ubuntu 10.04 LTS on hosts and quests, quest VLAN problem Date: Tue, 21 Dec 2010 21:16:01 +0100 (CET) Message-ID: <14019802.82.1292962973383.JavaMail.giedz@ipopema5019> References: <30930153.80.1292962125219.JavaMail.giedz@ipopema5019> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit To: kvm@vger.kernel.org Return-path: Received: from se91.internetdsl.tpnet.pl ([80.55.82.91]:36312 "EHLO mail.arise.pl" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751223Ab0LUUW7 (ORCPT ); Tue, 21 Dec 2010 15:22:59 -0500 Received: from localhost (localhost [127.0.0.1]) by mail.arise.pl (Postfix) with ESMTP id 5D88A389683 for ; Tue, 21 Dec 2010 21:16:02 +0100 (CET) Received: from mail.arise.pl ([127.0.0.1]) by localhost (mail1.arise [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 4yUYICosQNMr for ; Tue, 21 Dec 2010 21:16:01 +0100 (CET) Received: from mail1.arise (mail1.arise [192.168.199.7]) by mail.arise.pl (Postfix) with ESMTP id 76816385E3E for ; Tue, 21 Dec 2010 21:16:01 +0100 (CET) In-Reply-To: <30930153.80.1292962125219.JavaMail.giedz@ipopema5019> Sender: kvm-owner@vger.kernel.org List-ID: Hi, For the last few days we've been trying to set up ubuntu guest with vlans - until now no success. Here is the story: a) host is ubuntu 10.04 with one eth interface and bridge on it... host is called bdc1 # The primary network interface auto eth0 iface eth0 inet manual auto br0 iface br0 inet static # Static assign the IP, netmask, default gateway address 192.168.89.99 netmask 255.255.255.0 gateway 192.168.89.1 dns-nameservers 192.168.89.4 # Bind one or more interfaces to the bridge bridge_ports eth0 # Tune the bridge for a single interface bridge_stp on bridge_fd 0 bridge_maxwait 0 b) guest is bridged to br0 and it's called dns1 with the following xml conf file: dns1 79c9080e-65df-cdc5-2973-2c5ea97db872 1048576 1048576 2 hvm destroy restart restart /usr/bin/kvm c) on the guest network interfaces are as follows: auto eth0 iface eth0 inet static address 192.168.89.180 netmask 255.255.255.0 gateway 192.168.89.1 auto eth0.87 --->>> VLAN 87 -> vconfig add eth0 87 iface eth0.87 inet static address 192.168.87.4 netmask 255.255.255.0 d) guest's both interfaces are up, + switches are configured to handle VLAN87 + VLAN1 on the port "the host" is connected to -> we use HP ProCurve e) let ping this eth0.87 address 1st try from 192.168.87.2 to 192.168.87.4 sysadmin@dns1:~$ sudo tshark -i eth0.87 3.457098 192.168.87.2 -> 192.168.87.4 ICMP Echo (ping) request 3.457125 192.168.87.4 -> 192.168.87.2 ICMP Echo (ping) reply 3.458939 192.168.87.2 -> 192.168.87.4 ICMP Echo (ping) request 3.458957 192.168.87.4 -> 192.168.87.2 ICMP Echo (ping) reply 3.460546 192.168.87.2 -> 192.168.87.4 ICMP Echo (ping) request 3.460563 192.168.87.4 -> 192.168.87.2 ICMP Echo (ping) reply 3.462261 192.168.87.2 -> 192.168.87.4 ICMP Echo (ping) request 3.462278 192.168.87.4 -> 192.168.87.2 ICMP Echo (ping) reply 3.463659 192.168.87.2 -> 192.168.87.4 ICMP Echo (ping) request 3.463676 192.168.87.4 -> 192.168.87.2 ICMP Echo (ping) reply looks fine 2nd try from 192.168.87.4 (eth0.87) to 192.168.87.2 sysadmin@dns1:~$ ping 192.168.87.2 PING 192.168.87.2 (192.168.87.2) 56(84) bytes of data. 64 bytes from 192.168.87.2: icmp_seq=1 ttl=255 time=0.619 ms 64 bytes from 192.168.87.2: icmp_seq=2 ttl=255 time=0.670 ms 3rd try from "the host" 192.168.89.99 to 192.168.87.2 sysadmin@dns1:~$ sudo tshark -i eth0.87 -f 'icmp' Running as user "root" and group "root". This could be dangerous. Capturing on eth0.87 0.000000 192.168.89.99 -> 192.168.87.4 ICMP Echo (ping) request 0.999936 192.168.89.99 -> 192.168.87.4 ICMP Echo (ping) request 1.999826 192.168.89.99 -> 192.168.87.4 ICMP Echo (ping) request 2.999800 192.168.89.99 -> 192.168.87.4 ICMP Echo (ping) request 3.999770 192.168.89.99 -> 192.168.87.4 ICMP Echo (ping) request upsss I can see request BUT no replay!!!! f) I thought that maybe forward is not set but sysadmin@dns1:~$ cat /proc/sys/net/ipv4/ip_forward 1 sysadmin@bdc1:~$ cat /proc/sys/net/ipv4/ip_forward 1 g) of course pings between VLAN1 (.89.0) work great. What can be the problem? Do we miss something like VLAN should be set on host not on guest or something else... please help ;) Many thanks, Marcin