Can't create instance, no valid host was found, nova-compute is not running

asked 2014-11-25 08:58:43 -0600

Sliter gravatar image

I installed openstack via devstack, branch juno on my Ubuntu 12.04(not VM).

When I was trying to launch an instance, I received the error message "No valid host was found. There are not enough hosts available".

After googling a while, I found that someone has encountered exactly the same problem in this (thread).

Although the thread did not come with a final solution, it mentions that the missing of nova-compute could result in such a problem. And running the command "nova-manage service list" shows that nova-compute is indeed missing.

So I was trying to enable the nova-compute by the command "nova-manage service enable --host=xxx --service=nova-compute". However, I ended with an error message "error: Could not find binary nova-compute on host xxx."

Now, I am stuck here. Could anyone shed some light on this? Thanks

edit retag flag offensive close merge delete

8 answers

Sort by » oldest newest most voted

answered 2015-04-21 13:42:43 -0600

badmadrad gravatar image

My issue was because I did not have enough ram allocated to nova-scheduler on my bare-metal server. I wasn't afraid of overcommitment since it was dev machine.


sudo vim /etc/nova/nova.conf

Change: ram_allocation_ratio=1.5 to ram_allocation_ratio=3


sudo service openstack-nova-scheduler restart
sudo service openstack-nova-conductor restart
edit flag offensive delete link more

answered 2015-03-11 05:47:11 -0600

binou gravatar image

i got the same problem there is no nova compute in nova service and the file /var/log/nova/nova-compute.log does not exit

how can i fix it? can i add the nova-compute service without reinstalling all?

edit flag offensive delete link more


You have to look for that file in compute node. If that is not there, I suspect that there is something wrong with your installation.

bingoarunprasath gravatar imagebingoarunprasath ( 2015-03-19 08:00:40 -0600 )edit

i finally found the log in /opt/stack here are the logs and it seems that it is a libvirt version issue how can i fix it please?

binou gravatar imagebinou ( 2015-03-19 10:17:51 -0600 )edit

here is the seems that it is a libvirt version issue,isn't it?

service.start() File "/opt/stack/nova/nova/", line 164, in start self.manager.init_host() File "/opt/stack/nova/nova/compute/", line 1189, in init_host self.driver.init_host( File

binou gravatar imagebinou ( 2015-03-19 10:20:09 -0600 )edit

self._version_to_string(MIN_LIBVIRT_VERSION)) NovaException: Nova requires libvirt version 0.9.11 or greater. 2015-03-10 13:19:42.379 ERROR nova.openstack.common.threadgroup [req-9433abde-136a-482c-bd93-1d7220150edf None None] Nova requires libvirt version 0.9.11 or greater. 2015-03-10 13:19:42.379

binou gravatar imagebinou ( 2015-03-19 10:20:35 -0600 )edit

2015-03-10 13:19:42.379 TRACE nova.openstack.common.threadgroup self._version_to_string(MIN_LIBVIRT_VERSION)) 2015-03-10 13:19:42.379 TRACE nova.openstack.common.threadgroup NovaException: Nova requires libvirt version 0.9.11 or greater. 2015-03-10 13:19:42.379 TRACE nova.openstack.common.threadgrou

binou gravatar imagebinou ( 2015-03-19 10:20:53 -0600 )edit

answered 2014-12-06 07:07:09 -0600

This is a common error when the nova-scheduler can't find any suitable compute node to launch instance.

1) Check the output of 'nova-manage service list'. If 'nova-compute' and 'libvirt-bin' services are working fine, then it should show enabled and status should be :-)

2) If any compute node is available and enabled, check /var/log/nova/nova-compute.log in that node and see if you could find any errors. (see the old logs , as you cant see any logs while launching instance as the the request would be stopped at the scheduler itself)

3) If everything is fine and clean, then the compute node should be lack of resources. In this case, you can create a small flavor and try to launch instance with that flavor.

edit flag offensive delete link more

answered 2015-08-14 00:21:45 -0600

Suresh Dharavath gravatar image

updated 2015-08-14 06:56:04 -0600

It is mostly because you don't have the resources properly. Always try to use the small images rather than big ones and select the flavor m1.nano, m1.tiny( again it depends on how much ram and swap memories are available, In my case my swap space used to be 100% utilized and ram used to be 83% to 87% ). Don't go for flavors like medium, large, etc.., which causes the errors like what you mentioned.

edit flag offensive delete link more

answered 2015-05-21 01:22:39 -0600

检查logs目录下的出错信息,我的原因是BIOS没有开启VT,kvm模块未加载 (

edit flag offensive delete link more

answered 2016-05-20 01:53:32 -0600

GG1982 gravatar image

Hello! I got the same issue, and restarting the service nova-scheduler solves the case.

edit flag offensive delete link more

answered 2016-03-08 22:42:02 -0600

Before powering on the Instance, just tail logs. ( tail -f /var/log/nova/* | grep -i error ). so whenever you will receive error while creating instance, you can troubleshoot easily with the error, which is self explanatory for most of them. This procedure has helped me fix many of the issues.

edit flag offensive delete link more

answered 2018-04-19 04:59:41 -0600

syedkazmi gravatar image

I solved it by changing "Quota" for my "Demo" project of "myUser".

I had the same problem. I have installed OpenStack Training Labs. It has Project named Demo in which I was working. I had only 1Gb of RAM as Quota for that project. I changed the limit to 10Gb and I was able to launch the instance.

To this you have to be logged in as Admin (or have admin rights): Identity -> Projects -> Demo (you project name)

Hope it helps! :-)

edit flag offensive delete link more

Get to know Ask OpenStack

Resources for moderators

Question Tools



Asked: 2014-11-25 08:57:57 -0600

Seen: 59,396 times

Last updated: Mar 08 '16