Stein on rhel7.6: No valid host was found.There are not enough hosts available. [closed]
After creating 13 nodes by heat template, the 14th node throws the error "No valid host was found. There are not enough hosts available. Code: 500"
824 2019-06-05 13:58:47.946 7100 WARNING nova.scheduler.utils [req-0673d8fa-dc10-4167-ae3d-e7c54165dc87 ad73023f983a49e1a5c38e4ba0 b8d116 76c7bf1ad0dc4a5696fcf8a5400cad6e - default default] Failed to compute_task_build_instances: No valid host was found. Th ere are not enough hosts available.
825 Traceback (most recent call last):
826
827 File "/usr/lib/python2.7/site-packages/oslo_messaging/rpc/server.py", line 229, in inner
828 return func(*args, **kwargs)
829
830 File "/usr/lib/python2.7/site-packages/nova/scheduler/manager.py", line 168, in select_destinations
831 allocation_request_version, return_alternates)
832
833 File "/usr/lib/python2.7/site-packages/nova/scheduler/filter_scheduler.py", line 96, in select_destinations
834 allocation_request_version, return_alternates)
835
836 File "/usr/lib/python2.7/site-packages/nova/scheduler/filter_scheduler.py", line 265, in _schedule
837 claimed_instance_uuids)
838
839 File "/usr/lib/python2.7/site-packages/nova/scheduler/filter_scheduler.py", line 302, in _ensure_sufficient_hosts
840 raise exception.NoValidHost(reason=reason)
841
842 NoValidHost: No valid host was found. There are not enough hosts available.
843 : NoValidHost_Remote: No valid host was found. There are not enough hosts available.
844 2019-06-05 13:58:47.949 7100 WARNING nova.scheduler.utils [req-0673d8fa-dc10-4167-ae3d-e7c54165dc87 ad73023f983a49e1a5c38e4ba0 b8d116 76c7bf1ad0dc4a5696fcf8a5400cad6e - default default] [instance: 7acedfd0-8951-4ade-8f96-467c38b04f9c] Setting instance t o ERROR state.: NoValidHost_Remote: No valid host was found. There are not enough hosts available.
I think my resource is enough.
The output you share is not sufficient to understand why there is no suitable host. If debug logging is enabled, the scheduler tells you in its log which filter fails.
Hi, @Bernd Bausch I have found the reason. Because I divided the three OS::Nova::Server into one OS::Nova::ServerGroup by using scheduler_hints for anti-affinity, while I only have two compute nodes.
Solution: Add enough compute nodes, or reduce server in a group based on your compute nodes.