nova-compute service start stucks in a loop

asked 2018-11-25 06:43:03 -0500

soheil08 gravatar image

updated 2018-11-25 07:07:40 -0500


I'm installing OpenStack on two nodes. One as controller, network and compute, the other just as compute node. According to the document of Rocky, I installed all modules on the controller node and all became verified. At the next step, I configure the compute node and set the configurations according to the document, but when I start the service openstack-nova-compute it just logs:

compute.pps.wams systemd[1]: Starting OpenStack Nova Compute Server...

and it never start!

In the logs of nova-compute it just writes one line:

2018-11-25 15:53:56.730 1246 INFO os_vif [-] Loaded VIF plugins: ovs, linux_bridge, noop

I should tell that I don't config and install Neutron, yet because the documentation suggested installing components according to the list order. What could be the problem?

edit retag flag offensive close merge delete


Can you verify that no other nova-compute process is running? try ps alx | grep nova and kill respective process(es) if necessary. Try running the command manually /usr/bin/python /usr/bin/nova-compute and watch log files (nova-compute.log, syslog) for issues.

eblock gravatar imageeblock ( 2018-11-27 06:18:55 -0500 )edit

7 answers

Sort by ยป oldest newest most voted

answered 2018-12-05 02:03:40 -0500

soheil08 gravatar image

In my case, it was because of faulty Rabbitmq. Check the Rabbitmq logs and ensure it works properly.

edit flag offensive delete link more

answered 2019-04-02 19:28:01 -0500

Disable firewalld on the controller node

edit flag offensive delete link more

answered 2018-11-28 07:19:27 -0500

I have exactly the same problem and I can not find a solution anywhere. I hope someone shares useful information about it.


edit flag offensive delete link more



See the answer.

soheil08 gravatar imagesoheil08 ( 2018-12-05 02:04:16 -0500 )edit

answered 2019-03-10 15:58:00 -0500

LaurentDumont gravatar image

I have the same behavior following the deployment process for Rocky. It's a 4 node (controller + 3 compute) running Centos 7.

edit flag offensive delete link more

answered 2020-05-01 11:03:43 -0500

REEVZ gravatar image

Allowing port 5672 on controller's firewall solved this issue for me

  • on controller firewall-cmd --add-port=5672/tcp --permanent firewall-cmd --reload
edit flag offensive delete link more

answered 2019-01-19 22:38:54 -0500

Shellyyang gravatar image


I got the same problem. I checked the rabbitmq log and did not find errors. Is there alternative troubleshooting or solutions? Thanks.

Always stuck in 2019-01-19 23:20:34.074 16650 INFO os_vif [-] Loaded VIF plugins: ovs, linux_bridge, noop

edit flag offensive delete link more

answered 2019-01-28 15:58:24 -0500

Chuck Fouts gravatar image

I am also seeing the same issue. The last line in nova-compute.log is the same as reported. I've tried adding the compute node by following the (Rocky OpenStack instructions) and by trying to apply PackStack answer file with the exact same result. There isn't an error in the controller node's RabbitMQ logs and there are no discernable errors in any of the compute node's logs.

When running /usr/bin/nova-compute --debug the same behavior is observed.

edit flag offensive delete link more

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower


Asked: 2018-11-25 06:43:03 -0500

Seen: 1,128 times

Last updated: Dec 05 '18