Ask Your Question

med's profile - activity

2019-01-23 11:22:25 -0500 commented answer VNC Console in Dashboard fails to connect ot server (code: 1006)

Restarting the appropriate openstack-nova-compute (or nova-compute) and the various nova control node (nova-console, nova-api, but mostly nova-novncproxy) services after is important. In my case, the compute node's nova.conf was using hostnames that were not registered with the control node. IP +1.

2018-09-19 02:57:40 -0500 received badge  Good Answer (source)
2017-05-19 16:30:11 -0500 received badge  Nice Answer (source)
2017-01-30 13:46:17 -0500 received badge  Necromancer (source)
2016-10-03 20:48:09 -0500 answered a question What's the difference between shelving vs shutting down an instance?

The hypervisor (say libvirt for example) will still be aware of a VM in shutdown state. You can see this for yourself by doing a shutdown on an instance and then doing "virsh list --all" on the nova-compute (libvirt) node. You will see that there is a virsh domain defined for the shutdown instance. For a shelved instance, the instance is completely disconnected from the hypervisor so libvirt is no longer aware of the instance. Also there is no "contract" that the VM will restart when unshelved on the same hypervisor. (That's a good thing and one of the purposes of shelving.)

2015-07-14 12:52:56 -0500 answered a question libvirt broken cpuset on migration workarounds?

It's not clear to me what else you are omitting if you completely comment out that vcpu.set command. This should be an upstream bug against nova. Checking...

2015-04-08 19:27:12 -0500 received badge  Famous Question (source)
2015-01-15 16:31:45 -0500 received badge  Notable Question (source)
2015-01-15 00:30:19 -0500 received badge  Popular Question (source)
2015-01-14 22:49:19 -0500 received badge  Nice Question (source)
2015-01-14 17:28:44 -0500 received badge  Student (source)
2015-01-14 16:43:05 -0500 asked a question How do I nova boot from image create volume and specify the cinder backend to use?

nova boot --block-device seems to lack a way to boot from an image and create a new volume on a SPECIFIC CINDER BACKEND device. How do I do that? I have 3 different cinder backends.

2014-11-14 14:24:16 -0500 received badge  Enthusiast
2014-11-12 21:23:01 -0500 answered a question cinder-manage db sync Error

I've seen this as well in Trusty with Icehouse python-oslo.messaging and, here's the weird part, it works on some puppet hosts and not on others (all of which were installed with puppet ie, the same way.)

2014-09-09 01:50:41 -0500 received badge  Famous Question (source)
2013-11-06 01:29:31 -0500 received badge  Teacher (source)
2013-11-05 10:07:14 -0500 answered a question What is the role played by openstack in live migration?

OpenStack and cloud in general is more than just a hypervisor. OpenStack provides the infrastructure to move not only the VM instance (Nova), but also the underlying storage (Cinder) and the networking infrastructure (Neutron) so that the live migrated instance is a full complete instance throughout the migration and works just as if it has never been moved. Additionally, you can live migrate even if you do not have an underlying shared filesystem via block migration (along with VM migration.)

All of this is detailed in (Configuring Migrations)

2013-08-07 03:27:26 -0500 received badge  Notable Question (source)
2013-06-29 13:20:58 -0500 received badge  Popular Question (source)
2013-06-21 14:12:41 -0500 asked a question Can glance be setup as a dropbox

Can glance be setup as a dropbox so that third parties (say distributions such as Ubuntu, Red Hat, CentOS) can place their images into glance with simple or no authentication (ie, without an account on the glance server/openstack/keystone)?

(The images would likely be put into a "wip" box until they are reviewed/etc and only be made public after they were vetted.)

2012-05-17 20:16:52 -0500 answered a question Nova+Quantum+Openvswitch, two physical hosts, create VM, error


I'll reiterate what Somik says, this is NOT a normal configuration.

Take a look at what devstack recommends and understand that your second compute node can be much simpler than the first.

You might also want to refer to the Admin Guide and one of the 3rd party install guides (maybe this one: