Ask Your Question

pann1588's profile - activity

2014-07-17 10:50:57 -0600 received badge  Taxonomist
2014-03-08 04:54:44 -0600 received badge  Famous Question (source)
2014-03-08 04:54:44 -0600 received badge  Notable Question (source)
2013-12-27 01:42:24 -0600 received badge  Famous Question (source)
2013-12-24 02:05:18 -0600 received badge  Popular Question (source)
2013-12-17 19:31:57 -0600 commented question Neutron qrouter default route problem

Problem solved. It looks that the external net CIDR address pool I set had some issue. The original one I used is and the reason I selected it is that I tried to avoid any address collision with the existing hosts in the external net. After I changed to, the default routing entry was created successfully.

2013-12-16 23:12:32 -0600 commented question Neutron qrouter default route problem

Hi larsks, Thanks a lot. I've just updated the above post to show you related information.

2013-12-16 21:03:53 -0600 received badge  Editor (source)
2013-12-16 19:55:20 -0600 asked a question Neutron qrouter default route problem

Hi everyone,

We have a OpenStack deployment of three nodes: controller, network node, and compute node. After installing everything, I created external network and subnet in the "admin" user, and then login with tenant user to create tenant network and subnet. Then I created a router to interconnect the two subnets.

The I tried to ping outside from the VMs we launched but it didn't succeed. I traced ICMP packets and found that it can ping through both two interfaces (internal and external) of the router I created, but not to the external network. Then I used "sudo ip netns exec qroute-xxxx route -n " and found that the default route entry is missing. After I manually added the default route, then the ping went through succssfully. After associating VMs with floating IP, ping the VMs from outside were also successfully.

So, my question is, what is the reason for it, and which module is responsible for configuring the default routing table? What should I do to overcome such issue? Thanks a lot!

Regards, Jianli

Hi larsks,

Thanks a lot for the help. I use the dashboard to create all the networks and didn't use the command line. Shall I do that instead? I followed a tutorial written in Chinese on how to operate dashboard to create external and internal network. The link is: There are screenshots in this tutorial so hopefully you could get a sense of what it is trying to teach. I basically create an external network in admin account, and then login with customer tenant account to create internal network, and then created a router to connect the internal with external.

#neutron net-list: +--------------------------------------+---------+-------------------------------------------------------+ | id | name | subnets | +--------------------------------------+---------+-------------------------------------------------------+ | 69032459-4b0b-49d5-8088-31886c7faa9c | ext-net | 25760bbe-424f-4309-aa89-2afd7541c700 | | e5f83f9c-e633-4dd5-8021-23a870331784 | conaNet | 4feee8dd-77c4-4602-bdd5-d4121481da5c | +--------------------------------------+---------+-------------------------------------------------------+

#neutron router-list: +--------------------------------------+-------------+-----------------------------------------------------------------------------+ | id | name | external_gateway_info | +--------------------------------------+-------------+-----------------------------------------------------------------------------+ | ff25c07b-02ca-41e4-9299-e86593d3fc79 | cona_router | {"network_id": "69032459-4b0b-49d5-8088-31886c7faa9c", "enable_snat": true} | +--------------------------------------+-------------+-----------------------------------------------------------------------------+

#neutron router-port-list : +--------------------------------------+------+-------------------+--------------------------------------------------------------------------------------+ | id | name | mac_address | fixed_ips | +--------------------------------------+------+-------------------+--------------------------------------------------------------------------------------+ | 3bf353a2-b3f7-4ff0-b37c-09e1b0f75108 | | fa:16:3e:f3:1f:ec | {"subnet_id": "4feee8dd-77c4-4602-bdd5-d4121481da5c", "ip_address": ""} | | 5ad09e95-4c9e-4480-95a5-fbf7c2e289cd | | fa:16:3e:37:94:0c | {"subnet_id": "25760bbe-424f-4309-aa89-2afd7541c700", "ip_address": ""} | +--------------------------------------+------+-------------------+--------------------------------------------------------------------------------------+

Sorry for the formatting when I pasted it here.

Thanks again! Jianli

2013-12-12 00:16:19 -0600 received badge  Notable Question (source)
2013-12-11 01:28:55 -0600 received badge  Popular Question (source)
2013-12-10 23:00:09 -0600 commented question Nova-compute installation issue with Ubuntu 12.10

Thanks! The problem is solved when we switched it to Ubuntu 12.04 LTS.

2013-12-10 16:02:34 -0600 asked a question Nova-compute installation issue with Ubuntu 12.10


I have a three-nodes deployment of OpenStack using Havana in Ubuntu 12.10. I successfully installed controller and network node. However, when I tried to install the "Nova-compute-kvm and python-guestfs" services on the compute node, I run into trouble.

Because the "# add-apt-repository cloud-archive:havana" shown in the official tutorials simply do not work for me (maybe because of mine is Ubuntu 12.10, instead of 12.04 LTS? ) with resonse "Error: 'cloud-archive:havana' invalid". I added the repository manually by editing the "/etc/apt/sources.list" and adding "deb precise-updates/havana main". After running "apt-get install -y ubuntu-cloud-keyring," "apt-get update," and "apt-get upgrade" as I have done in both controller node and networking node, I tried to install the "Nova-compute-kvm and python-guestfs" services on the compute node. However, I run into the unmet dependency errors:

The following packages have unmet dependencies: nova-compute-kvm : Depends: libvirt-bin but it is not going to be installed Depends: python-libvirt but it is not going to be installed E: Unable to correct problems, you have held broken packages.

I tried this on another host and the same problem is reported. So I am wondering if it is the problem with Ubuntu 12.10, or with the havana repository. I googled a lot and cannot find useful information. Shall I replace this compute node with Ubuntu 12.04? Will it work OK with my controller and network node with Ubuntu 12.10 installed?

Does anybody have any clue? Thanks in advance!

Regards, Jianli