Ask Your Question
0

Newton Launching a new Instance. - "No valid host was found. There are not enough hosts available." [closed]

asked 2017-04-06 10:21:32 -0600

updated 2017-04-07 16:46:12 -0600

I am running Ubuntu 16.04 and OpenStack Newton. I have one server running as controller node.

When I am trying to launch new instance using CLI or horizon dashboard. An error occurs

No valid host was found. There are not enough hosts available.

I was trying to change ram and vCPU allocation ratio but it doesn't seems like it have any effect on it.

I think this might be something wrong with my Nova configuration. Here are my logs:

Error from horizon dashboard http://paste.openstack.org/show/605628/

This is what I see when I am issuing command nova hypervisor-statsIt seems like it does not see any available resources. Which I should have enough for at least small flavor instance. http://paste.openstack.org/show/605629/

However my nova quota-showproves I have plenty of resources to use http://paste.openstack.org/show/605632/

Here is my nova.conf file http://paste.openstack.org/show/605630/ and errors from nova-conductor.log http://paste.openstack.org/show/605649/

Also nova-scheduler.log file http://paste.openstack.org/show/605647/

This is also propably realted to the issue when I am accessing "Admin/System/All Hypervisors" via horizon dashboard the list is empty. I am not sure if I should see anything there?

edit:

This is what probably causing an issue no hypervisors available in Nova, still I do not know how to solve this problem.

nova hypervisor-list http://paste.openstack.org/show/605781/

edit retag flag offensive reopen merge delete

Closed for the following reason the question is answered, right answer was accepted by student
close date 2017-04-07 17:18:41.920900

Comments

1

Can you check the nova-scheduler log?

kuro gravatar imagekuro ( 2017-04-06 11:30:41 -0600 )edit

Here is my nova-scheduler.log file: http://paste.openstack.org/show/605647/

and errors from nova-conductor.log http://paste.openstack.org/show/605649/

student gravatar imagestudent ( 2017-04-06 12:45:43 -0600 )edit

4 answers

Sort by ยป oldest newest most voted
1

answered 2017-04-07 08:50:11 -0600

luowei gravatar image

updated 2017-04-07 08:58:21 -0600

First check nova.conf on controller and computes for lingering mismatched region names (i.e. "RegionOne" everywhere, not "openstacl" or "regionOne").

I don't think it is linked to allocation ratio, but I found I needed to explicitly provide the scheduler filters when specifying a ratio in nova.conf:

scheduler_default_filters =  ComputeFilter, CoreFilter, RamFilter, SameHostFilter
ram_allocation_ratio = 1.5
cpu_allocation_ratio = 30.0
edit flag offensive delete link more

Comments

You are right it is not allocation ratio it is no hypervisors available. There is nothing in the nova hypervisor list -> http://paste.openstack.org/show/605781/

But I do not know what can cause an issue

student gravatar imagestudent ( 2017-04-07 15:48:03 -0600 )edit
1

answered 2017-04-07 06:18:16 -0600

hkominos gravatar image

The logs look normal to me. The key line is Filter RetryFilter returned 0 hosts. Which means that there is no compute host with enough resources for your Vm. I would suggest.

1)shutdown all the compute nodes (from UI) and keep only 1 (or force through availability zones) 2)Check the nova and neutron logs on the compute nodes! to see what they report as status and what happens after they receive the request.

edit flag offensive delete link more

Comments

I only have one device - controller node (no compute nodes) but I did find what is an issue. However I do not know how to fix it - Nova do not have any hypervisors, the table is empty. This is what it shows: http://paste.openstack.org/show/605781/

student gravatar imagestudent ( 2017-04-07 15:46:40 -0600 )edit

So that means that you have no compute nodes. You need compute nodes to spawn your Vms in.

hkominos gravatar imagehkominos ( 2017-04-10 05:54:29 -0600 )edit
0

answered 2017-04-06 12:57:51 -0600

juangallego gravatar image

I took a look at your logs, and I can tell you that I had the same issue installing manually What I did was to reinstall using packstack and it worked without any problem. My guess is that it was a configuration issue, maybe you want to try devstack or packstack to generate the configuration files you need.

edit flag offensive delete link more

Comments

I would like to avoid using ready solutions, especially when I am that close to actually running it. This is the last thing which is stopping me from running it.

student gravatar imagestudent ( 2017-04-06 13:47:59 -0600 )edit
0

answered 2017-04-07 17:17:34 -0600

I found the problem it was /etc/nova/nova-compute.conf file where I had some random charachters. I think it was my configuration problem. I have different problem now.

edit flag offensive delete link more

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower

Stats

Asked: 2017-04-06 10:21:32 -0600

Seen: 651 times

Last updated: Apr 07 '17