Hi, I continue my tests with Magnum Caracal and found another strange behaviour during cluster creation. It goes through state CREATION_FAILED before reaching the state CREATION_COMPLETE (HEALTHY). It is very confusing. Looking at Heat engine logs, it happens during load balancer creation, at the time the LB amphora has been created and Octavia is waiting to be able to connect it (typically one minue). During this period, Heat logs the following error: ----------- 2024-08-21 09:54:52.194 3200318 INFO heat.engine.resource [None req-610b24bd-7511-4d31-9f8c-b922162a7648 jouvin - - - - -] CREATE: LoadBalancer "loadbalancer" [670985ef-5808-4226-ad1b-b1244871de16] Stack "test-mj-1-30-m1n1-bis-ipufjomqmnnf-etcd_lb-d5fuqygvpp2j" [5c5905bf-f538-4a42-93ce-38d96b7c6603] 2024-08-21 09:54:52.194 3200318 ERROR heat.engine.resource Traceback (most recent call last): 2024-08-21 09:54:52.194 3200318 ERROR heat.engine.resource File "/usr/lib/python3.9/site-packages/heat/engine/resource.py", line 922, in _action_recorder 2024-08-21 09:54:52.194 3200318 ERROR heat.engine.resource yield 2024-08-21 09:54:52.194 3200318 ERROR heat.engine.resource File "/usr/lib/python3.9/site-packages/heat/engine/resource.py", line 1034, in _do_action 2024-08-21 09:54:52.194 3200318 ERROR heat.engine.resource yield from self.action_handler_task(action, args=handler_args) 2024-08-21 09:54:52.194 3200318 ERROR heat.engine.resource File "/usr/lib/python3.9/site-packages/heat/engine/resource.py", line 984, in action_handler_task 2024-08-21 09:54:52.194 3200318 ERROR heat.engine.resource done = check(handler_data) 2024-08-21 09:54:52.194 3200318 ERROR heat.engine.resource File "/usr/lib/python3.9/site-packages/heat/engine/resources/openstack/octavia/loadbalancer.py", line 169, in check_create_complete 2024-08-21 09:54:52.194 3200318 ERROR heat.engine.resource return self._check_status() 2024-08-21 09:54:52.194 3200318 ERROR heat.engine.resource File "/usr/lib/python3.9/site-packages/heat/engine/resources/openstack/octavia/octavia_base.py", line 29, in _check_status 2024-08-21 09:54:52.194 3200318 ERROR heat.engine.resource raise exception.ResourceInError(resource_status=status) 2024-08-21 09:54:52.194 3200318 ERROR heat.engine.resource heat.common.exception.ResourceInError: Went to status ERROR due to "Unknown" 2024-08-21 09:54:52.194 3200318 ERROR heat.engine.resource 2024-08-21 09:54:58.804 3200319 INFO heat.engine.scheduler [None req-610b24bd-7511-4d31-9f8c-b922162a7648 jouvin - - - - -] Task pause timed out 2024-08-21 09:54:59.340 3200318 INFO heat.engine.scheduler [None req-610b24bd-7511-4d31-9f8c-b922162a7648 jouvin - - - - -] Task pause timed out ----- that results in Magnum setting the state to CREATE_FAILED despite it is a transient state from Heat point of view and despite Heat stack creation continues successfully once the transient state is cleared and Magnum later updates the status to CREATE_COMPLETED. Unfortunately magnum-conductor log has nothing related to these events (it is filled with barbican_client related messages mainly). Any idea what could cause this behaviour? Best regards, Michel