No valid host was found
I delete a lot of created instances but now i cannot create any new instance any more because of this error:
{u'message': u'No valid host was found. ', u'code': 500, u'created': u'2019-01-22T15:14:31Z'}
I tink the deleted instances they still exist in the database, so i have to delete them.
How can i delete openstack instances from database ?
/var/log/nova/nova-scheduler.log
2019-01-23 11:27:46.166 15296 ERROR nova.scheduler.client.report [req-e8476dcb-6e6b-40de-bc49-686407e1714b 29ae9e5f770f42db9e62638112e43bd2 25811a9477684c0a8d89a58e84d116c6 - default default] Failed to retrieve allocation candidates from placement API for filters: RequestGroup(use_same_provider=False, resources={DISK_GB:1, MEMORY_MB:64, VCPU:1}, traits=[], aggregates=[]) ` 2019-01-23 11:27:46.167 15296 INFO nova.scheduler.manager [req-e8476dcb-6e6b-40de-bc49-686407e1714b 29ae9e5f770f42db9e62638112e43bd2 25811a9477684c0a8d89a58e84d116c6 - default default] Got no allocation candidates from the Placement API. This could be due to insufficient resources or a temporary occurrence as compute nodes start up.
The database format can change from release to release. There are no universal instructions.
First dump the database and confirm that they really exist. They shouldn’t.
Also check the scheduler log for the real cause of the problem.
The nova database is exist.
In the nova-scheduler.log i found these errors:
Failed to retrieve allocation candidates from placement api for filters. This could be due to insufficient resources or a temporary occurrence as compute nodes start up.
More details above in the question description
@Bernd Bausch i use this command select * from compute_nodes \G;
ram_allocation_ratio: NULL cpu_allocation_ratio: NULL
is it normal result ?
There is something wrong with the Placement service. Install the openstack client Placement plugin if it's not installed yet, and look into it.
I don't know enough about the internal DB to answer the second question.