reworked bmc blueprint to fix bug in kubernetes docker-multinode scripts

EarthmanT · Apr 19, 2017 · e06553a · e06553a
1 parent 454aba4
commit e06553a
Show file tree

Hide file tree

Showing 3 changed files with 270 additions and 107 deletions.
diff --git a/bmc-blueprint.yaml b/bmc-blueprint.yaml
@@ -4,16 +4,16 @@ imports:
   - http://www.getcloudify.org/spec/cloudify/3.4/types.yaml
   - http://www.getcloudify.org/spec/fabric-plugin/1.3.1/plugin.yaml
   - http://www.getcloudify.org/spec/diamond-plugin/1.3.3/plugin.yaml
-  - https://github.com/cloudify-incubator/cloudify-oraclebmc-plugin/archive/0.1.zip
+  - https://raw.githubusercontent.com/cloudify-incubator/cloudify-oraclebmc-plugin/master/plugin.yaml
   - types/scale.yaml
 
 dsl_definitions:
   bmc_config: &bmc_config
     user: 
     fingerprint: 
-    key_file: ~/.ssh/bmcs_api_key
+    key_file: 
     tenancy: 
-    region: us-phoenix-1
+    region: 
 
   hyperkube_monitoring: &hyperkube_monitoring
     collectors_config:
@@ -23,6 +23,7 @@ dsl_definitions:
       DiskUsageCollector:
         config:
           devices: sd[a-z]+[0-9]*$
+      NetworkCollector: {}
       ProcessResourcesCollector:
         config:
           enabled: true
@@ -46,7 +47,6 @@ inputs:
     description: image (must be Oracle Linux)
   master_shape:
     description: flavor
-    default: BM.Standard1.36
   worker_image:
     description: image
   worker_shape:
@@ -69,22 +69,25 @@ node_templates:
     interfaces:
       cloudify.interfaces.lifecycle:
         start:
-          implementation: fabric.fabric_plugin.tasks.run_commands
+          implementation: fabric.fabric_plugin.tasks.run_task
           inputs:
             fabric_env:
               host_string: { get_attribute: [ master_host, public_ip ] }
               user: { get_input: ssh_user }
               key_filename: { get_input: ssh_keyfile }
-            commands:
-              - "curl -LO https://storage.googleapis.com/kubernetes-release/release/$(curl -s https://storage.googleapis.com/kubernetes-release/release/stable.txt)/bin/linux/amd64/kubectl"
-              - "chmod +x kubectl"
-              - "sudo setenforce 0"
-              - "sudo systemctl disable firewalld"
-              - "sudo systemctl stop firewalld"
-              - "sudo yum install -y git"
-              - "rm -rf kube-deploy"
-              - "git clone https://github.com/kubernetes/kube-deploy"
-              - "cd kube-deploy/docker-multinode;sudo ./master.sh"
+            tasks_file: scripts/fabric_tasks.py
+            task_name: start_master                
+            task_properties:
+              k8s_settings:
+                k8s_version: v1.3.0
+                etcd_version: 2.2.5
+                flannel_version: v0.6.2
+                flannel_ipmasq: 'true'
+                flannel_network: 10.1.0.0/16
+                flannel_backend: udp
+                restart_policy: unless-stopped
+                arch: amd64
+                net_interface: eth0
     relationships:
       - type: cloudify.relationships.contained_in
         target: master_host
@@ -94,17 +97,26 @@ node_templates:
     interfaces:
       cloudify.interfaces.lifecycle:
         start:
-          implementation: fabric.fabric_plugin.tasks.run_commands
+          implementation: fabric.fabric_plugin.tasks.run_task
           inputs:
             fabric_env:
               host_string: { get_attribute: [ worker_host, public_ip ] }
               user: { get_input: ssh_user }
               key_filename: { get_input: ssh_keyfile }
-            commands:
-              - "sudo yum install -y git"
-              - "rm -rf kube-deploy"
-              - "git clone https://github.com/kubernetes/kube-deploy"
-              - { concat: [ "cd kube-deploy/docker-multinode;sudo MASTER_IP=", { get_attribute: [ master_host, private_ip ] }," ./worker.sh" ] }
+            tasks_file: scripts/fabric_tasks.py
+            task_name: start_worker
+            task_properties:
+              master_ip: { get_attribute: [ master_host, ip ] }
+              k8s_settings:
+                k8s_version: v1.3.0
+                etcd_version: 2.2.5
+                flannel_version: v0.6.2
+                flannel_ipmasq: 'true'
+                flannel_network: 10.1.0.0/16
+                flannel_backend: udp
+                restart_policy: unless-stopped
+                arch: amd64
+                net_interface: eth0
     relationships:
       - type: cloudify.relationships.depends_on
         target: master
@@ -119,10 +131,10 @@ node_templates:
       bmc_config: *bmc_config
       ssh_keyfile: { get_input: master_key}
       name: master
-      public_key_file: ~/.ssh/dfilppi-dc.key.pub
+      public_key_file: 
       image_id: { get_input: master_image }
       instance_shape: { get_input: master_shape }
-      compartment_id: ''
+      compartment_id: 
       availability_domain: { get_input: availability_domain }
     relationships:
       - type: cloudify.oraclebmc.relationships.instance_connected_to_subnet
@@ -136,10 +148,10 @@ node_templates:
       bmc_config: *bmc_config
       ssh_keyfile: { get_input: worker_key}
       name: worker
-      public_key_file: ~/.ssh/dfilppi-dc.key.pub
+      public_key_file: 
       image_id: { get_input: worker_image }
       instance_shape: { get_input: worker_shape }
-      compartment_id: ''
+      compartment_id: 
       availability_domain: { get_input: availability_domain }
     relationships:
       - type: cloudify.oraclebmc.relationships.instance_connected_to_subnet
@@ -165,16 +177,16 @@ node_templates:
     properties:
       bmc_config: *bmc_config
       use_external_resource: true
-      resource_id: ''
+      resource_id: 
 
   subnet:
     type: cloudify.oraclebmc.nodes.Subnet
     properties:
       bmc_config: *bmc_config
       name: kubernetes_subnet
-      compartment_id: ''
+      compartment_id: 
       cidr_block: 10.10.20.0/24
-      availability_domain: vNkz:PHX-AD-2
+      availability_domain: 
       security_rules: 
         - "0.0.0.0/0,22"
         - "0.0.0.0/0,53"
@@ -194,93 +206,91 @@ node_templates:
   gateway:
     type: cloudify.oraclebmc.nodes.Gateway
     properties:
-      resource_id: ''
+      resource_id: 
       use_external_resource: true
       bmc_config: *bmc_config
     relationships:
       - type: cloudify.oraclebmc.relationships.gateway_connected_to_network
         target: network
 
-
 groups:
-
- scale_up_group:
-   members: [worker_host]
-    # This defines a scale group whose members may be scaled up, incrementing by 1.
+
+  scale_up_group:
+    members: [worker_host]
+     # This defines a scale group whose members may be scaled up, incrementing by 1.
+     # The scale worflow is called when the following criteria are met
+     # The Hyperkube process total CPU will be more than 3 for a total of 10 seconds.
+     # No more than 6 hosts will be allowed.
+    policies:
+      auto_scale_up:
+        type: scale_policy_type
+        properties:
+          policy_operates_on_group: true
+          scale_limit: 6
+          scale_direction: '<'
+          scale_threshold: 30
+          service_selector: .*worker_host.*cpu.total.user
+          cooldown_time: 60
+        triggers:
+          execute_scale_workflow:
+            type: cloudify.policies.triggers.execute_workflow
+            parameters:
+              workflow: scale
+              workflow_parameters:
+                delta: 1
+                scalable_entity_name: worker
+                scale_compute: true
+
+  scale_down_group:
+    # This defines a scale group whose members may be scaled down. Only one host will be removed per run.
     # The scale worflow is called when the following criteria are met
-    # The Hyperkube process total CPU will be more than 3 for a total of 10 seconds.
-    # No more than 6 hosts will be allowed.
-   policies:
-     auto_scale_up:
-       type: scale_policy_type
-       properties:
-         policy_operates_on_group: true
-         scale_limit: 6
-         scale_direction: '<'
-         scale_threshold: 30
-         #service_selector: .*worker_host.*.cpu.total.user
-         service_selector: .*worker_host.*cpu.total.user
-         cooldown_time: 60
-       triggers:
-         execute_scale_workflow:
-           type: cloudify.policies.triggers.execute_workflow
-           parameters:
-             workflow: scale
-             workflow_parameters:
-               delta: 1
-               scalable_entity_name: worker
-               scale_compute: true
-
- scale_down_group:
-   # This defines a scale group whose members may be scaled down. Only one host will be removed per run.
-   # The scale worflow is called when the following criteria are met
-   # The Hyperkube process total CPU will be less than 1 for a total of 200 seconds.
-   # No less than 2 hosts will be allowed.
-   members: [worker_host]
-   policies:
-     auto_scale_down:
-       type: scale_policy_type
-       properties:
-         scale_limit: 2
-         scale_direction: '>'
-         scale_threshold: 25
-         #service_selector: .*worker_host.*.process.hyperkube.cpu.percent
-         service_selector: .*worker_host.*cpu.total.user
-         cooldown_time: 60
-         moving_window_size: 30
-       triggers:
-         execute_scale_workflow:
-           type: cloudify.policies.triggers.execute_workflow
-           parameters:
-             workflow: scale
-             workflow_parameters:
-               delta: -1
-               scalable_entity_name: worker
-               scale_compute: true
-
- heal_group:
-   # This defines a group of hosts in members that may be healed.
-   # The heal workflow is called when a the following policy criteria are met.
-   # Either the hyperkube process on the host, or the total host CPU need fall silent.
-   # The host and all software that it is supposed to have running on it will be healed.
-   members: [worker_host]
-   policies:
-     simple_autoheal_policy:
-       type: cloudify.policies.types.host_failure
-       properties:
-         service:
-           - .*worker_host.*.cpu.total.system
-           - .*worker_host.*.process.hyperkube.cpu.percent
-         interval_between_workflows: 60
-       triggers:
-         auto_heal_trigger:
-           type: cloudify.policies.triggers.execute_workflow
-           parameters:
-             workflow: heal
-             workflow_parameters:
-               node_instance_id: { 'get_property': [ SELF, node_id ] }
-               diagnose_value: { 'get_property': [ SELF, diagnose ] }
-
+    # The Hyperkube process total CPU will be less than 1 for a total of 200 seconds.
+    # No less than 2 hosts will be allowed.
+    members: [worker_host]
+    policies:
+      auto_scale_down:
+        type: scale_policy_type
+        properties:
+          scale_limit: 2
+          scale_direction: '>'
+          scale_threshold: 25
+          #service_selector: .*worker_host.*.process.hyperkube.cpu.percent
+          service_selector: .*worker_host.*cpu.total.user
+          cooldown_time: 60
+          moving_window_size: 30
+        triggers:
+          execute_scale_workflow:
+            type: cloudify.policies.triggers.execute_workflow
+            parameters:
+              workflow: scale
+              workflow_parameters:
+                delta: -1
+                scalable_entity_name: worker
+                scale_compute: true
+
+  heal_group:
+    # This defines a group of hosts in members that may be healed.
+    # The heal workflow is called when a the following policy criteria are met.
+    # Either the hyperkube process on the host, or the total host CPU need fall silent.
+    # The host and all software that it is supposed to have running on it will be healed.
+    members: [worker_host]
+    policies:
+      simple_autoheal_policy:
+        type: cloudify.policies.types.host_failure
+        properties:
+          service:
+            - .*worker_host.*.cpu.total.system
+            - .*worker_host.*.process.hyperkube.cpu.percent
+          interval_between_workflows: 60
+        triggers:
+          auto_heal_trigger:
+            type: cloudify.policies.triggers.execute_workflow
+            parameters:
+              workflow: heal
+              workflow_parameters:
+                node_instance_id: { 'get_property': [ SELF, node_id ] }
+                diagnose_value: { 'get_property': [ SELF, diagnose ] }
+
 outputs:
   kubernetes_info:
     description: Kubernetes Dashboard URL

diff --git a/resources/kube-deploy.tgz b/resources/kube-deploy.tgz