When changing parameters in an Instance, e.g. BROKERS_COUNT: "4"
for Kafka or FLINK_TASKMANAGER_REPLICAS: "3"
for Flink I would expect Maestro to scale up/down accordingly.
Nothing really happens, though I see some (?unrelated) error messages as well as information about plans that could lead us on the right path.
2019/01/28 21:02:28 Error getting FrameworkVersion flink-1.7 for instance flink: FrameworkVersion.maestro.k8s.io "flink-1.7" not found
maestro-demo $ kubectl get frameworkversions
NAME CREATED AT
flink-1.7 8m
2019/01/28 21:02:28 Recieved create event for &{{Instance maestro.k8s.io/v1alpha1} {flink default /apis/maestro.k8s.io/v1alpha1/namespaces/default/instances/flink 56e2f7a8-2382-11e9-822f-42010a800154 2949 1 2019-01-28 20:57:12 -0800 PST <nil> <nil> map[framework:flink controller-tools.k8s.io:1.0] map[kubectl.kubernetes.io/last-applied-configuration:{"apiVersion":"maestro.k8s.io/v1alpha1","kind":"Instance","metadata":{"annotations":{},"labels":{"controller-tools.k8s.io":"1.0","framework":"flink"},"name":"flink","namespace":"default"},"spec":{"frameworkVersion":{"name":"flink-1.7","namespace":"default","type":"FrameworkVersion"},"parameters":{"FLINK_TASKMANAGER_REPLICAS":"3"}}}
] [] nil [] } {{ default flink-1.7 } [] map[FLINK_TASKMANAGER_REPLICAS:3]} {{ default flink-deploy-997290000 f4bf9ccc-2382-11e9-822f-42010a800154 } COMPLETE}}
2019/01/28 21:02:28 Error getting FrameworkVersion flink-1.7 for instance flink: FrameworkVersion.maestro.k8s.io "flink-1.7" not found
2019/01/28 21:02:28 Adding flink-deploy-997290000 to reconcile
2019/01/28 21:02:28 Adding flink-deploy-997290000 to reconcile
{"level":"info","ts":1548738148.9631062,"logger":"kubebuilder.controller","caller":"controller/controller.go:134","msg":"Starting Controller","Controller":"instance-controller"}
{"level":"info","ts":1548738148.963074,"logger":"kubebuilder.controller","caller":"controller/controller.go:134","msg":"Starting Controller","Controller":"planexecution-controller"}
{"level":"info","ts":1548738148.9630651,"logger":"kubebuilder.controller","caller":"controller/controller.go:134","msg":"Starting Controller","Controller":"framework-controller"}
{"level":"info","ts":1548738148.9630919,"logger":"kubebuilder.controller","caller":"controller/controller.go:134","msg":"Starting Controller","Controller":"frameworkversion-controller"}
{"level":"info","ts":1548738149.0641012,"logger":"kubebuilder.controller","caller":"controller/controller.go:153","msg":"Starting workers","Controller":"planexecution-controller","WorkerCount":1}
2019/01/28 21:02:29 PlanExecution flink-deploy-243510000 has already run to completion, not processing.
2019/01/28 21:02:29 PlanExecution flink-deploy-997290000 has already run to completion, not processing.
{"level":"info","ts":1548738149.0677109,"logger":"kubebuilder.controller","caller":"controller/controller.go:153","msg":"Starting workers","Controller":"instance-controller","WorkerCount":1}
{"level":"info","ts":1548738149.067785,"logger":"kubebuilder.controller","caller":"controller/controller.go:153","msg":"Starting workers","Controller":"frameworkversion-controller","WorkerCount":1}
{"level":"info","ts":1548738149.0677068,"logger":"kubebuilder.controller","caller":"controller/controller.go:153","msg":"Starting workers","Controller":"framework-controller","WorkerCount":1}
2019/01/28 21:02:29 FrameworkController: Recieved Reconcile request for flink
2019/01/28 21:02:29 FrameworkVersionController: Recieved Reconcile request for flink-1.7
2019/01/28 21:02:40 InstanceController: UpdateInstance: Going to call plan deploy
2019/01/28 21:02:40 Current Plan for Instance is already done, wont change the Suspend flag
2019/01/28 21:02:40 InstanceController: Recieved Reconcile request for flink
2019/01/28 21:02:40 Old and new spec matched...
2019/01/28 21:02:40 InstanceController: UpdateInstance: Going to call plan
2019/01/28 21:02:40 InstanceController: Recieved Reconcile request for flink
2019/01/28 21:02:40 Phase 0 Step 0 has 3 objects
2019/01/28 21:02:40 CreateOrUpdate resulted in: unchanged
2019/01/28 21:02:40 Unkonwn type is marked healthy by default
2019/01/28 21:02:40 CreateOrUpdate resulted in: unchanged
2019/01/28 21:02:40 Unkonwn type is marked healthy by default
2019/01/28 21:02:40 CreateOrUpdate resulted in: unchanged
2019/01/28 21:02:40 Unkonwn type is marked healthy by default
2019/01/28 21:02:40 Phase flink has strategy serial
2019/01/28 21:02:40 Phase flink marked as serial
2019/01/28 21:02:40 Step jobmanager is healthy, so I can continue on
2019/01/28 21:02:40 Step jobmanager looked at
2019/01/28 21:02:40 Phase flink is healthy
2019/01/28 21:02:40 Phase flink marked as healthy
2019/01/28 21:02:40 Phase flink is healthy
2019/01/28 21:02:40 PlanExecution flink-deploy-28047000 has already run to completion, not processing.