Ask Your Question

Problems with Instances getting IP

asked 2013-11-12 09:02:20 -0500

Mike77 gravatar image

Good Morning,

 I am a total newb with Openstack so I have many questions, and alot are being answered as I bulldoze my way through standing up my own stack and playing with it. I have 4 nodes setup. One controller with 3 Compute nodes. Right now my biggest problem I am having is getting the new Instances that I create to pull in IP addresses through DHCP. When I see the instance in the Openstack Controller page, they are given IP addresses and I can do a nova list on the Compute node and they show up with IP addresses, but whenever I attempt to restart my CentOS minimal instances, they get a failed error message. SO I guess I am at a loss on how it can show that I have an IP address but yet I cannot seem to get my actual instance to get that ip address.

When I look at the logs on the Instance itself, I am given the following info.

DHCP DISCOVER on eth0 to - port 67
No DHCP OFFER Recieved

I am guessing this is my DHCP request being broadcasted out on port 67 to get an address, but no DHCP server is giving an address? Which is my DHCP server? Is it the controller?

After reviewing the logs on the compute nodes, I did some more digging and found these errors... 


 neutron:skipping unknown group key: Firewall-driver

 I did do some research on my own with trusty Google and found lots of different answers to the various issues... but some of them seem like they are rather complex. So before I get too carried away here, I wanted to bounce some of this off some openstack gurus and see if there was any pointers they could offer. How is the basic networking model set up? Have you seen anything like this before? Also noted that there was some info about quantum services. I can see several different quantum related scripts, but no master startup script, or what I am guessing is this quantum-agent script.

Any help would be greatly apprecaited!

edit retag flag offensive close merge delete


Are you using quantum/neutron or nova-network?

simone.tiraboschi gravatar imagesimone.tiraboschi ( 2013-11-12 09:18:01 -0500 )edit

2 answers

Sort by ยป oldest newest most voted

answered 2013-11-12 12:05:52 -0500

dheeru gravatar image

Hi 1. You are hitting the DHCP issue and it is originating from DNSMASQ. 2. Can you run 'ps -aef | grep dnsmasq' on controller ? Does it show that it is running ? 3. There is way to get the dhcp lease from command line. This is to verify that DHCP is working perfectly. I forgot the command. You can search. Else I will give after some time. 4. Can you verify things according to following link ? 5. Do you see an errros like AMQP server: [Errno 111] ECONNREFUSED in nova*.log or some other log files ?

edit flag offensive delete link more



Okay so I did a check for the ECONNREFUSED and I have yet to find it anywhere. Trying to find the commands to do a command line check for DHCP. DNSMASQ is running on the controller...

Mike77 gravatar imageMike77 ( 2013-11-12 13:32:42 -0500 )edit

Correcting the dnsmasq 2.49 error I was seeing in the logs. I was trying to install the latest, but however for CentOS it looks like that only 2.48 is the latest in the repo. Anyone know where i can get a new version of the dnsmasq and I am guessing dnsmasq-utils as well? Need to add to repo server

Mike77 gravatar imageMike77 ( 2013-11-12 16:12:36 -0500 )edit

Hi How is your issue now ? Is it resolved ? Can you try using the dhclient command ?

dheeru gravatar imagedheeru ( 2013-11-13 07:15:35 -0500 )edit

Well I did a couple of things. First I went ahead and reinstalled Openstack via RDO packstack without the neutron option as an all in one node. Thinking about trying to add more nodes down the road. So sort of running a devstack I guess.. but I am getting IP's to my VM, so that problem is solved.

Mike77 gravatar imageMike77 ( 2013-11-14 17:57:19 -0500 )edit

However and this might be a new one, I am having problems with volumes. Apparently when I create a volume and assign it to a VM, it gets truncated to an IDE or /dev/hdc device. I tell it to be /dev/vdc but it always changes. WHy is that? I have not found any real solution out there, yet

Mike77 gravatar imageMike77 ( 2013-11-14 17:58:28 -0500 )edit

answered 2013-11-12 09:54:03 -0500

Mike77 gravatar image

I believe neutron... I used the packstack script from RDO to do the setup, so whatver it set up is what I am using.

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


Asked: 2013-11-12 09:02:20 -0500

Seen: 2,664 times

Last updated: Nov 12 '13