Note: The message returned is generic and further troubleshooting must be done for the root cause.
There are several areas to check in 5.1:
- On the compute node where the instance was attempting to launch, review the /var/log/nova/nova-compute.log for errors related to the request id.
- There is a possibility all hosts were filtered in which we need to track the request ID or the instance ID in the /var/log/nova/nova-scheduler.log on the controllers.
- Check the services and components of VMware Integrated OpenStack for errors.
- Are there any ESXi host or storage failures?
- Is communication working with the NSX Manager or vCenter?
- Is the failure only happening with a specific message or metadata requirement?
In 6.0 or 7.x for VMware Integrated OpenStack:
- Collect the list of pods with: osctl get pods
- Find the request ID for the instance in the nova compute pod with: osctl log<compute pod UUID>
- Then search the nova-scheduler pod's logs the request, and find the result that invalidated available hosts: osctl log <nova-scheduler pod UUID>