Ask Your Question
0

multiple nics for the guest vm

asked 2012-06-22 23:29:10 -0500

Hello, I am running Essex released version in Quantum networking mode (using openvswitch). 1 controller - runs nova-compute/nova-network/nova-api/nova-volume/etc 1 compute node - runs nova-compute/nova-api/nova-api-metadataI tried launching Maverick image which has multiple eth nics set up. I used 3 nics (app-net, web-net and db-net) to launch it. I was able to launch it successfully. I got 3 fixed ip addresses from three networks. 0b4e1c8d-6d64-4625-a30e-a9b456455ab3 | app-server | ACTIVE | public-app-net=192.168.4.12; public-db-net=192.168.5.4; public-web-net=192.168.6.4 |

I am able to ping and ssh into the instance only thru 192.168.4.12. But, I am not able to ping/ssh thru 192.168.5.4 and 192.168.6.4. To work around this, I logged into the instance and in /etc/network/interfaces, I added auto eth1 iface eth1 inet dhcp auto eth2 iface eth2 inet dhcpand did sudo /etc/init.d/networking restart.

Then, I saw 192.168.5.4 and 192.168.6.4 addresses getting assigned to eth1 and eth2 of the guest vm. I was able to see the receiving and transmitting traffic on both gateways and tap interfaces. 3 pairs of dnsmasq processes are created , one for each network. However, I am still not able to ping or ssh the guest thru 192.168.5.4 and 192.168.6.4 from the host machine . Any lead is appreciated. Thanks, -jsb

edit retag flag offensive close merge delete

2 answers

Sort by » oldest newest most voted
0

answered 2012-06-22 23:39:39 -0500

salvatore-orlando gravatar image

Hi Jay,

at a first glance ( no pun - i.e. reference to your creature ;) - intended ) I don't see anything wrong with openstack configuration. The fact that you see 3 dnsmasq pairs lead me to believe there are actually three distinct bridges, which is fine. However, I don't recall whether nova's network managers associate IP addresses to the bridges on the host. If not, you probably won't have a route from the host to the networks where eth1 and eth2 are attached, which I assume to be /24. If this hypothesis is correct you might be able to reach eth0 because it's in the same IP subnet as the host, but the latter will not have information for reaching eth1 and eth2 subnets.

Regards, Salvatore

On 23 June 2012 00:31, jay question201225@answers.launchpad.net wrote:

New question #201225 on quantum: https://answers.launchpad.net/quantum/+question/201225 (https://answers.launchpad.net/quantum...)

Hello, I am running Essex released version in Quantum networking mode (using openvswitch). 1 controller - runs nova-compute/nova-network/nova-api/nova-volume/etc 1 compute node - runs nova-compute/nova-api/nova-api-metadataI tried launching Maverick image which has multiple eth nics set up. I used 3 nics (app-net, web-net and db-net) to launch it.  I was able to launch it successfully. I got 3 fixed ip addresses from three networks.  0b4e1c8d-6d64-4625-a30e-a9b456455ab3 | app-server | ACTIVE  | public-app-net=192.168.4.12; public-db-net=192.168.5.4; public-web-net=192.168.6.4 |

I am able to ping and ssh into the instance only thru 192.168.4.12. But, I am not able to ping/ssh thru 192.168.5.4 and 192.168.6.4. To work around this, I logged into the instance and in /etc/network/interfaces, I added auto eth1 iface eth1 inet dhcp auto eth2 iface eth2 inet dhcpand did sudo /etc/init.d/networking restart.

Then, I saw 192.168.5.4 and 192.168.6.4 addresses getting assigned to eth1 and eth2 of the guest vm. I was able to see the receiving and transmitting traffic on both gateways and tap interfaces. 3 pairs of dnsmasq processes are created , one for each network. However, I am still not able to ping or ssh the guest thru 192.168.5.4 and 192.168.6.4 from the host machine . Any lead is appreciated. Thanks, -jsb


You received this question notification because you are a member of Netstack Core Developers, which is an answer contact for quantum.

edit flag offensive delete link more
0

answered 2012-06-26 00:21:36 -0500

gongysh gravatar image

regarding you have to modify the interfaces file as you have done, In general image assumes it has only one card in interfaces file, openstack will not modify the interfaces file in dhcp mode, so you have to modify the file yourself. This is my opinion.

for cannot ssh, please show the br-int and br-tun with ovs-ofctl and ovs-vsctl also output of ifconfig -a and iptables-save and ebtables-save.

Thanks

edit flag offensive delete link more

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower

Stats

Asked: 2012-06-22 23:29:10 -0500

Seen: 48 times

Last updated: Jun 26 '12