[openstack-dev] [openstack-ansible] [nova] Compute node failing to come up in hypervisor list after first reboot

Lawrence J. Albinson lawrence at ljalbinson.com
Mon Mar 6 17:01:28 UTC 2017

Hey Andy,

Thanks again. Yes, it is almost certainly an SSL issue. I re-built with SSL turned off and things now work.

In the morning (UK time) I'm going to re-build with SSL on and add 'insecure=True' to the nova config. I'll report the result when done.

Kindests, Lawrence

From: Andy McCrae
Sent: 06 March 2017 15:14
To: OpenStack Development Mailing List (not for usage questions)
Subject: Re: [openstack-dev] [openstack-ansible] [nova] Compute node failing to come up in hypervisor list after first reboot

Hey Lawrence,

The nc -z is checking that you can connect to the port, but the error itself looks to be related to the handshake when using SSL certificates in rabbitmq.

"getrandom() initialization failed. (_ssl.c:590)" being the important bit there. You'll probably find more if you start searching around that area rather than the connection failure itself.

On a side note - this sounds like it may well be a bug with SSL certs in rabbitmq in OpenStack-Ansible, I'd love to get that fixed if that is the case!

Hope that helps though.


On 6 March 2017 at 15:38, Lawrence J. Albinson <lawrence at ljalbinson.com<mailto:lawrence at ljalbinson.com>> wrote:
I've been chasing a problems with a build for some days now and would greatly appreciate any pointers the community can offer.

The build is built of three nodes (infra+storage+compute) each of which is a KVM virtual machine.

After installation everything seems fine. However, after rebooting the compute node the hypervisor list reports it as being down.

The /var/log/nova/nova-compute.log is filling up with messages of the form:

     2017-03-06 14:13:11.209 4380 ERROR oslo.messaging._drivers.impl_rabbit [req-7fa52faa-c21f-4a1c-8413-fddad9e52c2e - - - - -] [fd4c494d-1ef5-4502-991f-521f0a8a2bc3] AMQP server on<> is unreachable: getrandom() initialization failed. (_ssl.c:590). Trying again in 32 seconds. Client port: None

I can confirm that is in fact reachable from compute1.

Furthermore,  'nc -z -v -w5 5671' reports:

    Connection to 5671 port [tcp/amqps] succeeded!

Any pointers would be greatly appreciated.

Kind regards, Lawrence

OpenStack Development Mailing List (not for usage questions)
Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe<http://OpenStack-dev-request@lists.openstack.org?subject:unsubscribe>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20170306/90164aaa/attachment.html>

More information about the OpenStack-dev mailing list