New nova compute nodes not correctly adding into service

Laurent Dumont laurentfdumont at gmail.com
Fri Dec 25 21:41:10 UTC 2020


The nova.filter log entry for the nova-scheduler should tell you why
computes are being filtered out.

You can also check the flavors you are using to see if it's targeting a
host-aggregate.

On Thu, Dec 24, 2020 at 11:40 AM Grant Morley <grant at civo.com> wrote:

> Hi All,
>
> I was wondering if anyone could point me in the right direction with an
> issue we are having adding in new compute nodes to our platform. We have
> been trying to add a few new compute nodes to increase our capacity and
> upon configuring them with OSA and them appearing to be in service we have
> noticed that no new instances are provisioning onto them. We did a test by
> disabling all compute nodes except for the new ones and an instance launch
> failed with "No valid compute hosts were found"
>
> We have enabled debug mode on Nova and also looked at the nova-api
> containers ( scheduler, compute etc ) and it appears there is not even an
> attempt to try and launch an instance on the new hosts from what we can see.
>
> The only thing we have noticed on the newly added hosts is that the
> "privsep-helper" doesn't appear to be started whereas on a working node it
> is running.
>
> We have not changed the software versions of OpenStack or OSA, we are
> running Queens.
>
> The newly added compute node is reporting itself back fine from what we
> can see:
>
>  Total usable vcpus: 72, total allocated vcpus: 0
> _report_final_resource_view
> /openstack/venvs/nova-17.1.2/lib/python2.7/site-packages/nova/compute/resource_tracker.py:827
> 2020-12-24 16:26:32.895 6491 INFO nova.compute.resource_tracker
> [req-6c58a97e-b7cb-4987-a219-da5b1ce156f8 - - - - -] Final resource view:
> name=compute-27.openstack.local phys_ram=257826MB used_ram=2048MB
> phys_disk=30166GB used_disk=0GB total_vcpus=72 used_vcpus=0 pci_stats=[]
> 2020-12-24 16:26:33.022 6491 DEBUG nova.compute.resource_tracker
> [req-6c58a97e-b7cb-4987-a219-da5b1ce156f8 - - - - -] Compute_service record
> updated for compute-27:compute-27.openstack.local
> _update_available_resource
> /openstack/venvs/nova-17.1.2/lib/python2.7/site-packages/nova/compute/resource_tracker.py:767
> 2020-12-24 16:26:33.023 6491 DEBUG oslo_concurrency.lockutils
> [req-6c58a97e-b7cb-4987-a219-da5b1ce156f8 - - - - -] Lock
> "compute_resources" released by
> "nova.compute.resource_tracker._update_available_resource" :: held 0.334s
> inner
> /openstack/venvs/nova-17.1.2/lib/python2.7/site-packages/oslo_concurrency/lockutils.py:285
> 2020-12-24 16:26:35.001 6491 DEBUG oslo_service.periodic_task
> [req-6c58a97e-b7cb-4987-a219-da5b1ce156f8 - - - - -] Running periodic task
> ComputeManager._poll_rebooting_instances run_periodic_tasks
> /openstack/venvs/nova-17.1.2/lib/python2.7/site-packages/oslo_service/periodic_task.py:215
>
> Does anyone know of any other logs we can check / enable that might lead
> us in the right direction to figure out why the new hosts are not adding in?
>
> Any help would be greatly appreciated as we are a bit stumped at the
> moment.
>
> Many thanks,
> --
> Grant Morley
> Site Reliability Engineer, Civo Ltd
> Unit H-K, Gateway 1000, Whittle Way
> Stevenage, Herts, SG1 2FP, UK
> Visit us at www.civo.com
> Signup for an account now > <https://www.civo.com/signup>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-discuss/attachments/20201225/2ad275b3/attachment.html>


More information about the openstack-discuss mailing list