Ask Your Question
2

Openstack fails to spawn instance!

asked 2017-11-08 14:41:12 -0500

jcorkey gravatar image

I have deployed openstack using kolla and kolla-ansible

Info about my environment:

  • OS: centos 7
  • kolla/kolla-ansible version: stable/pike
  • kolla image tags: 5.0.1
  • Nodes(deploy targets): one control node, one compute node

I can navigate to my horizon dashboard and things seem to be working when creating/adding routers, networks, flavors,security groups, etc. etc. but I can NOT launch an instance. I have been trying to figure out my problem for a while and so far no one has been able to help. Seems like most people don't have experience using kolla.

Here is information about my problem.

The horizon dashboards gives an ambiguous error saying Error: No valid host was found. There are not enough hosts available.]

The main error I see in the logs is in the neutron-server.log on my control node which I have listed below:

2017-11-08 18:35:24.458 24 ERROR neutron.plugins.ml2.managers [req-4874b31c-3d79-4a6a-b327-b7e73ede9c4c f9d20d0f9b724fcca5b290bb320401ac 06cc4bcfd6534694b0052b31ebd8fd28 - default default] Failed to bind port 14638c2a-3f0d-45d9-9e0d-a10ca7162835 on host openstackcompute.glv.internal for vnic_type normal using segments [{'network_id': '79f94edb-2ccf-4a2a-8aca-bfe68f2eb61d', 'segmentation_id': 39, 'physical_network': None, 'id': '0494fc8d-9310-434b-956b-eda660713eb8', 'network_type': u'vxlan'}]

I see this error print out a lot as soon as I try and launch an instance.

I have deployed openstack on my internal network so my deploy host, control node, and compute have no internet access.

The quick start guide in the link below says to have a network_interface and neutron_external_interface. docs.openstack.org/kolla-ansible/latest/user/quickstart.html

It says the neutron_external_interface allows instances to access the rest of the internet but because I have no internet on my servers I DID NOT set the neutron_external_interface to an up and running interface i.e. neutron_external_interface=em2 So I only have one NIC on both nodes

both control and compute nodes have one up and running interface named em1 the acts as the network_interface and they can ping each other.

Could only having one NIC network_interface be causing the error that is shown in the logs above or could it be something else?? Would I even need the neutron_external_interface since I don't want the nodes to have access to the rest of the internet?

Please help me.

edit retag flag offensive close merge delete

1 answer

Sort by ยป oldest newest most voted
0

answered 2017-11-09 08:42:33 -0500

timi gravatar image

anything from nova.log ? probably there are several configuration issues you have

edit flag offensive delete link more

Comments

Yea the error that appears in the nove.log on the controller is below:

jcorkey gravatar imagejcorkey ( 2017-11-09 11:53:50 -0500 )edit

oslo_service.periodic_task [req-b36aa07b-360d-4b1d-a71b-a1e68a648391 - - - - -] Error during SchedulerManager._expire_reservations: DBConnectionError: (pymysql.err.OperationalError) (2003, "Can't connect to MySQL server on 'IP_ADDRESS' ([Errno 113] EHOSTUNREACH)")

jcorkey gravatar imagejcorkey ( 2017-11-09 11:54:09 -0500 )edit

looks like the scheduler can't connect to mysql server, check you mysql settings

timi gravatar imagetimi ( 2017-11-10 08:02:21 -0500 )edit

Where exactly would I check those mysql settings? Since kolla deploys openstack services in docker containers, whats the retinue for modifying configurations and settings for those services running in docker containers? I not sure how to change configurations and then apply those configurations.

jcorkey gravatar imagejcorkey ( 2017-11-10 08:53:49 -0500 )edit

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: 2017-11-08 14:41:12 -0500

Seen: 170 times

Last updated: Nov 09 '17