Ask Your Question
0

Network configuration / Can't ping or ssh to our instances

asked 2011-02-17 23:58:17 -0500

everett-toews gravatar image

We have been trying to deploy nova (Bexar) for the past week on a single machine. We can install and configure nova on a fresh Ubuntu 10.10 server and even successfully start VMs but we have not been able to connect to those VMs. Neither ping nor ssh work (yes, we have done the euca-authorize steps).

We've tried all of the steps from the http://wiki.openstack.org/TroubleshootingNova (http://wiki.openstack.org/Troubleshoo...) . We've tried asking other questions, https://answers.launchpad.net/nova/+question/145062 (https://answers.launchpad.net/nova/+q...) and https://answers.launchpad.net/nova/+question/145063 (https://answers.launchpad.net/nova/+q...) (no replies). We've tried going on IRC, giving bits and pieces of information about our problems and getting good troubleshooting tips in return but, ultimately, nothing that has worked for us. We've read about a similar problems: https://answers.launchpad.net/nova/+question/141960 (https://answers.launchpad.net/nova/+q...) was already solved and the patch merged in Bexar and https://answers.launchpad.net/nova/+question/145567 (https://answers.launchpad.net/nova/+q...) doesn't seem to be relevant anymore (can't find --flat_interface reference in any of the nova- --help).

So, out of desperation, here is all of the information about our configuration (IP addresses changed) in the hopes that someone can pinpoint what the problem is.

Is our networking configured properly for Nova? Why can't we ping or ssh to our instances?

Any answers or network troubleshooting tips are greatly appreciated!

** Installation Architecture ** By design our network architecture very closely resembles the Nova example installation architecture.

http://docs.openstack.org/openstack-compute/admin/content/figures/NOVA_install_arch.png (http://docs.openstack.org/openstack-c...)

Keep in mind that, at the moment, we are only trying to install to a single machine.

** Installation Procedure ** We are doing the scripted installation ( http://docs.openstack.org/openstack-compute/admin/content/ch03s02.html#d5e161 (http://docs.openstack.org/openstack-c...) ). When we run nova-CC-install-v1.1.sh here are the values we input:

S3 Host IP (Default is 99.99.99.185 -- ENTER to accept): RabbitMQ Host IP (Default is 99.99.99.185 -- ENTER to accept): Cloud Controller Host IP (Default is 99.99.99.185 -- ENTER to accept): mySQL Host IP (Default is 99.99.99.185 -- ENTER to accept):

Controller network range for ALL projects (normally x.x.x.x/12):192.168.0.0/16 Total amount of usable IPs for ALL projects:64 Nova project user name:admin Nova project name:admin-project Desired network + CIDR for project (normally x.x.x.x/24):192.168.0.0/24 How many networks for project:1 How many availible IPs per project network:8

Please enter your local server IP (Default is 99.99.99.185 -- ENTER to accept): Please enter your broadcast IP (Default is 99.99.99.191 -- ENTER to accept): Please enter your netmask (Default is 255.255.255.224 -- ENTER to accept): Please enter your gateway (Default is 99 ... (more)

edit retag flag offensive close merge delete

11 answers

Sort by » oldest newest most voted
0

answered 2011-07-11 12:27:40 -0500

the error message of console log is below:

2011-07-11 10:04:11,496 - DataSourceEc2.py[WARNING]: waiting for metadata service at http://169.254.169.254/2009-04-04/meta-data/instance-id (http://169.254.169.254/2009-04-04/met...)

2011-07-11 10:04:11,498 - DataSourceEc2.py[WARNING]: 10:04:11 [ 1/100]: url error [[Errno 113] No route to host]

edit flag offensive delete link more
0

answered 2011-07-11 12:22:50 -0500

moreover, we have one dhcp and gateway :10.0.0.1 and are not allow to have another dhcp server since they will conflict in the same network.

how do I config openstack's network to avoid this issue in my case ?

edit flag offensive delete link more
0

answered 2011-07-11 12:11:21 -0500

I meet the same issue in the single node installation with one NIC enabled. how do I fix this issue? thanks a lot, Further, I have only 10.x.x.x subnet, not 192.168.x.x subnet.

here is my nova.conf:

--dhcpbridge_flagfile=/etc/nova/nova.conf --dhcpbridge=/usr/bin/nova-dhcpbridge --logdir=/var/log/nova --state_path=/var/lib/nova --lock_path=/var/lock/nova --verbose --s3_host=10.101.1.142 --rabbit_host=10.101.1.142 --cc_host=10.101.1.142 --ec2_url=http://10.101.1.142:8773/services/Cloud --fixed_range=192.168.0.0/16 --network_size=64 --FAKE_subdomain=ec2 --routing_source_ip=10.101.1.142 --verbose --sql_connection=mysql://root:iforgot@10.101.1.142/nova --network_manager=nova.network.manager.FlatManager

edit flag offensive delete link more
0

answered 2011-03-14 16:01:19 -0500

everett-toews gravatar image

Hi Sunny,

Sorry I missed your question until now. If you haven't solved your problem already, I would advise starting a new question as this one has be solved already (in my case). It's best to have separate questions for these types of problems. Even though the symptoms of the problem are similar the answer may be very different.

Thanks, Everett

edit flag offensive delete link more
0

answered 2011-02-25 03:54:35 -0500

I am redoing the entire steps again as mentioned in - -I am seeing some issues 0: http://docs.openstack.org/openstack-compute/admin/content/ch03s02.html (http://docs.openstack.org/openstack-c...)

1:

Installing packages ################### Installing package 'python-software-properties' ... Already installed Installing package 'python-mysqldb' ... Already installed Installing package 'mysql-server' ... Already installed Installing package 'nova-api' ... ok Installing package 'nova-network' ... ok Installing package 'nova-objectstore' ... Already installed Installing package 'nova-scheduler' ... ok Installing package 'nova-compute' ... ok Installing package 'unzip' ... Already installed Installing package 'vim' ... Already installed Installing package 'euca2ools' ... Already installed Finalizing mySQL setup ######################

ERROR 1045 (28000): Access denied for user 'root'@'localhost' (using password: YES) ERROR 1045 (28000): Access denied for user 'root'@'localhost' (using password: YES) ERROR 1045 (28000): Access denied for user 'root'@'localhost' (using password: YES) ...done...

Notice the errors

3:the script took long time at enabling SSH to running instances...lot of time - -over 30 mins etc

4: Notice I am trying to connect to Amazon EC2 as my Cloud Controller + S3

Please help me find the root of the issue reported in #2 and why delay in #3.

Will this Script work - if my Cloud Controller and S3 EC2_URL is AMAZON EC2? I have all the required info like EC2 CERT EC2 Secrete Key Access Key Access Secret key

edit flag offensive delete link more
0

answered 2011-02-24 22:53:05 -0500

Can you tell me which installation process you had followed? 1: NovaInstall/DevInstallScript?? 2:NovaInstall/DevPkgInstall?? 3:NovaInstall/Bexar??

I tried Bexar and DevInstallScript -- not sure when went wrong - I could not connect.

When tried to start nova-xyz services -- they dont throw up errors. However - no logs are produced + every time I try to do "ps -aef | grep nova" -- I am seeing different process IDs -- this means something wrong.

Could you please help me?

edit flag offensive delete link more
0

answered 2011-02-24 22:31:15 -0500

Hi Everett Toews I am seeing the same problem as your. I am glad to hear that your is resolved

Can you help me?

edit flag offensive delete link more
0

answered 2011-02-18 14:35:11 -0500

everett-toews gravatar image

Definitely willing to give it a try but I've never accessed a VM via the serial console before. How would I go about doing that?

We're using KVM by the way.

Thanks, Everett

On Feb 18, 2011, at 5:13 AM, Christian Berendt question145820@answers.launchpad.net wrote:

Your question #145820 on OpenStack Compute (nova) changed: https://answers.launchpad.net/nova/+question/145820 (https://answers.launchpad.net/nova/+q...)

Christian Berendt posted a new comment: connect to the running instance through the serial console and look if the network in the instance is coming up.


You received this question notification because you are a direct subscriber of the question.

edit flag offensive delete link more
0

answered 2011-02-23 16:47:04 -0500

everett-toews gravatar image

On a suggestion from Jordan, I switched to FlatDHCPManager for the networking mode and that fixed this problem.

To /etc/nova/nova.conf I added: --network_manager=nova.network.manager.FlatDHCPManager --flat_network_dhcp_start=192.168.0.2 --public_interface=eth0 --flat_interface=eth1 --flat_injected=False

I also removed the bridge from /etc/network/interfaces and restarted networking: /etc/init.d/networking restart

Then I restarted all of the nova services: restart libvirt-bin restart nova-network restart nova-compute restart nova-api restart nova-objectstore restart nova-scheduler

edit flag offensive delete link more
0

answered 2011-02-18 12:13:12 -0500

berendt gravatar image

connect to the running instance through the serial console and look if the network in the instance is coming up.

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: 2011-02-17 23:58:17 -0500

Seen: 137 times

Last updated: Jul 11 '11