HAproxy LBaaS loadbalancer failled to update by AutoScalingGroup
Most times during stack creation from HOT template (autoscalinggroup with instance and lb pollmember) heat was not able to add an instance to loadbalancer pool because loadbalancer object is in UPDATE_PENDING state. Stack create fails with error:
Create_Failed: Resource CREATE failed: StateInvalidClient: resources.lb_cluster.resources.xd6y6bcknxku.resources.lb_node: Invalid state PENDING_UPDATE of loadbalancer resource dc139962-379c-46e9-a445-b2dcfcfebbaf Neutron server returns request_ids: ['req-1a9c3c53-ac7a-4f5b-8efc-9c06709313a7']
There is 2 of 3 members in a pool - first and third.
Same situation with stack deletion
Delete_Failed: Resource DELETE failed: StateInvalidClient: resources.lb_cluster.resources.n3vl2nx2yn6x.resources.lb_node: Invalid state PENDING_UPDATE of loadbalancer resource 897b0939-1d15-47e6-80b8-d4f8f9fc920f Neutron server returns request_ids: ['req-80329d45-8e1a-4643-94e8-fe0244ff371d']
But usually from next few attempts creation or deletion works. Looks like requests on load balancer update sometimes takes too long and sequential request fails.
Any suggestions? Thanks
I'd recommend to start with enable debug/verbose and inspect lbass' / heat's logs
Please raise a bug at https://bugs.launchpad.net/heat