2i2c-aws-us: Nodegroup split and k8s update

config/clusters/2i2c-aws-us/cosmicds.values.yaml

@@ -38,6 +38,8 @@ jupyterhub:
           name: Cosmic DS, Harvard
           url: https://www.cosmicds.cfa.harvard.edu/
   singleuser:
+    nodeSelector:
+      2i2c/hub-name: cosmicds
     # No persistent storage should be kept to reduce any potential data
     # retention & privacy issues.
     # Ref https://github.com/2i2c-org/infrastructure/issues/2128#issuecomment-1635107926

config/clusters/2i2c-aws-us/dask-staging.values.yaml

@@ -28,6 +28,8 @@ basehub:
             name: 2i2c
             url: https://2i2c.org
     singleuser:
+      nodeSelector:
+        2i2c/hub-name: dask-staging
       image:
         name: pangeo/pangeo-notebook
         tag: "latest"
@@ -39,3 +41,14 @@ basehub:
           authenticator_class: "github"
         GitHubOAuthenticator:
           oauth_callback_url: "https://dask-staging.aws.2i2c.cloud/hub/oauth_callback"
+
+dask-gateway:
+  gateway:
+    scheduler:
+      extraPodConfig:
+        nodeSelector:
+          2i2c/hub-name: dask-staging
+    worker:
+      extraPodConfig:
+        nodeSelector:
+          2i2c/hub-name: dask-staging

config/clusters/2i2c-aws-us/itcoocean.values.yaml

@@ -45,6 +45,8 @@ jupyterhub:
         admin_users:
           - eeholmes # Eli Holmes, Community representative
   singleuser:
+    nodeSelector:
+      2i2c/hub-name: itcoocean
     # Requested in https://2i2c.freshdesk.com/a/tickets/1320
     defaultUrl: /lab
     # shared-public for collaboration

config/clusters/2i2c-aws-us/ncar-cisl.values.yaml

@@ -45,6 +45,8 @@ basehub:
             - NicholasCote # Nicholas Cote, Initial administrator
             - nwehrheim # Nick Wehrheim, Community representative
     singleuser:
+      nodeSelector:
+        2i2c/hub-name: ncar-cisl
       image:
         # image choice preliminary and is expected to be setup via
         # https://ncar-cisl.2i2c.cloud/services/configurator/ by the community
@@ -250,3 +252,14 @@ basehub:
               node.kubernetes.io/instance-type: g4dn.xlarge
             extra_resource_limits:
               nvidia.com/gpu: "1"
+
+dask-gateway:
+  gateway:
+    scheduler:
+      extraPodConfig:
+        nodeSelector:
+          2i2c/hub-name: ncar-cisl
+    worker:
+      extraPodConfig:
+        nodeSelector:
+          2i2c/hub-name: ncar-cisl

config/clusters/2i2c-aws-us/showcase.values.yaml

@@ -49,6 +49,8 @@ basehub:
         Authenticator:
           enable_auth_state: true
     singleuser:
+      nodeSelector:
+        2i2c/hub-name: showcase
       extraEnv:
         SCRATCH_BUCKET: s3://2i2c-aws-us-scratch-showcase/$(JUPYTERHUB_USER)
         PERSISTENT_BUCKET: s3://2i2c-aws-us-persistent-showcase/$(JUPYTERHUB_USER)
@@ -246,3 +248,14 @@ basehub:
               node.kubernetes.io/instance-type: g4dn.xlarge
             extra_resource_limits:
               nvidia.com/gpu: "1"
+
+dask-gateway:
+  gateway:
+    scheduler:
+      extraPodConfig:
+        nodeSelector:
+          2i2c/hub-name: showcase
+    worker:
+      extraPodConfig:
+        nodeSelector:
+          2i2c/hub-name: showcase

config/clusters/2i2c-aws-us/staging.values.yaml

@@ -33,5 +33,7 @@ jupyterhub:
       GitHubOAuthenticator:
         oauth_callback_url: "https://staging.aws.2i2c.cloud/hub/oauth_callback"
   singleuser:
+    nodeSelector:
+      2i2c/hub-name: staging
     extraEnv:
       SCRATCH_BUCKET: s3://2i2c-aws-us-scratch-staging/$(JUPYTERHUB_USER)

eksctl/2i2c-aws-us.jsonnet

@@ -25,13 +25,106 @@ local nodeAz = "us-west-2a";
 // A `node.kubernetes.io/instance-type label is added, so pods
 // can request a particular kind of node with a nodeSelector
 local notebookNodes = [
-    { instanceType: "r5.xlarge" },
-    { instanceType: "r5.4xlarge" },
-    { instanceType: "r5.16xlarge" },
+    // staging
+    {
+        instanceType: "r5.xlarge",
+        namePrefix: "nb-staging",
+        labels+: { "2i2c/hub-name": "staging" },
+        tags+: { "2i2c:hub-name": "staging" }
+    },
+    {
+        instanceType: "r5.4xlarge",
+        namePrefix: "nb-staging",
+        labels+: { "2i2c/hub-name": "staging" },
+        tags+: { "2i2c:hub-name": "staging" }
+    },
+    {
+        instanceType: "r5.16xlarge",
+        namePrefix: "nb-staging",
+        labels+: { "2i2c/hub-name": "staging" },
+        tags+: { "2i2c:hub-name": "staging" }
+    },
+    // dask-staging
+    {
+        instanceType: "r5.xlarge",
+        namePrefix: "nb-dask-staging",
+        labels+: { "2i2c/hub-name": "dask-staging" },
+        tags+: { "2i2c:hub-name": "dask-staging" }
+    },
+    {
+        instanceType: "r5.4xlarge",
+        namePrefix: "nb-dask-staging",
+        labels+: { "2i2c/hub-name": "dask-staging" },
+        tags+: { "2i2c:hub-name": "dask-staging" }
+    },
+    {
+        instanceType: "r5.16xlarge",
+        namePrefix: "nb-dask-staging",
+        labels+: { "2i2c/hub-name": "dask-staging" },
+        tags+: { "2i2c:hub-name": "dask-staging" }
+    },
+    // showcase
+    {
+        instanceType: "r5.xlarge",
+        namePrefix: "nb-showcase",
+        labels+: { "2i2c/hub-name": "showcase" },
+        tags+: { "2i2c:hub-name": "showcase" }
+    },
+    {
+        instanceType: "r5.4xlarge",
+        namePrefix: "nb-showcase",
+        labels+: { "2i2c/hub-name": "showcase" },
+        tags+: { "2i2c:hub-name": "showcase" }
+    },
+    {
+        instanceType: "r5.16xlarge",
+        namePrefix: "nb-showcase",
+        labels+: { "2i2c/hub-name": "showcase" },
+        tags+: { "2i2c:hub-name": "showcase" }
+    },
+    {
+        instanceType: "g4dn.xlarge",
+        namePrefix: "nb-showcase",
+        minSize: 0,
+        labels+: { "2i2c/hub-name": "showcase" },
+        tags+: {
+            "k8s.io/cluster-autoscaler/node-template/resources/nvidia.com/gpu": "1",
+            "2i2c:hub-name": "showcase",
+        },
+        taints+: {
+            "nvidia.com/gpu": "present:NoSchedule"
+        },
+        // Allow provisioning GPUs across all AZs, to prevent situation where all
+        // GPUs in a single AZ are in use and no new nodes can be spawned
+        availabilityZones: masterAzs,
+    },
+    // ncar-cisl
+    {
+        instanceType: "r5.xlarge",
+        namePrefix: "nb-ncar-cisl",
+        labels+: { "2i2c/hub-name": "ncar-cisl" },
+        tags+: { "2i2c:hub-name": "ncar-cisl" }
+    },
+    {
+        instanceType: "r5.4xlarge",
+        namePrefix: "nb-ncar-cisl",
+        labels+: { "2i2c/hub-name": "ncar-cisl" },
+        tags+: { "2i2c:hub-name": "ncar-cisl" }
+    },
+    {
+        instanceType: "r5.16xlarge",
+        namePrefix: "nb-ncar-cisl",
+        labels+: { "2i2c/hub-name": "ncar-cisl" },
+        tags+: { "2i2c:hub-name": "ncar-cisl" }
+    },
     {
         instanceType: "g4dn.xlarge",
+        namePrefix: "nb-ncar-cisl",
+        minSize: 0,
+        labels+: { "2i2c/hub-name": "ncar-cisl" },
         tags+: {
-            "k8s.io/cluster-autoscaler/node-template/resources/nvidia.com/gpu": "1"
+            "k8s.io/cluster-autoscaler/node-template/resources/nvidia.com/gpu": "1",
+            "2i2c:hub-name": "ncar-cisl",
         },
         taints+: {
             "nvidia.com/gpu": "present:NoSchedule"
@@ -40,6 +133,44 @@ local notebookNodes = [
         // GPUs in a single AZ are in use and no new nodes can be spawned
         availabilityZones: masterAzs,
     },
+    // itcoocean
+    {
+        instanceType: "r5.xlarge",
+        namePrefix: "itcoocean",
+        labels+: { "2i2c/hub-name": "itcoocean" },
+        tags+: { "2i2c:hub-name": "itcoocean" }
+    },
+    {
+        instanceType: "r5.4xlarge",
+        namePrefix: "itcoocean",
+        labels+: { "2i2c/hub-name": "itcoocean" },
+        tags+: { "2i2c:hub-name": "itcoocean" }
+    },
+    {
+        instanceType: "r5.16xlarge",
+        namePrefix: "itcoocean",
+        labels+: { "2i2c/hub-name": "itcoocean" },
+        tags+: { "2i2c:hub-name": "itcoocean" }
+    },
+    // cosmicds
+    {
+        instanceType: "r5.xlarge",
+        namePrefix: "cosmicds",
+        labels+: { "2i2c/hub-name": "cosmicds" },
+        tags+: { "2i2c:hub-name": "cosmicds" }
+    },
+    {
+        instanceType: "r5.4xlarge",
+        namePrefix: "cosmicds",
+        labels+: { "2i2c/hub-name": "cosmicds" },
+        tags+: { "2i2c:hub-name": "cosmicds" }
+    },
+    {
+        instanceType: "r5.16xlarge",
+        namePrefix: "cosmicds",
+        labels+: { "2i2c/hub-name": "cosmicds" },
+        tags+: { "2i2c:hub-name": "cosmicds" }
+    },
 ];
 
 
@@ -54,7 +185,24 @@ local daskNodes = [
     // A not yet fully established policy is being developed about using a single
     // node pool, see https://github.com/2i2c-org/infrastructure/issues/2687.
     //
-    { instancesDistribution+: { instanceTypes: ["r5.4xlarge"] }},
+    {
+        namePrefix: "dask-staging",
+        labels+: { "2i2c/hub-name": "dask-staging" },
+        tags+: { "2i2c:hub-name": "dask-staging" },
+        instancesDistribution+: { instanceTypes: ["r5.4xlarge"] }
+    },
+    {
+        namePrefix: "dask-showcase",
+        labels+: { "2i2c/hub-name": "showcase" },
+        tags+: { "2i2c:hub-name": "showcase" },
+        instancesDistribution+: { instanceTypes: ["r5.4xlarge"] }
+    },
+    {
+        namePrefix: "dask-ncar-cisl",
+        labels+: { "2i2c/hub-name": "ncar-cisl" },
+        tags+: { "2i2c:hub-name": "ncar-cisl" },
+        instancesDistribution+: { instanceTypes: ["r5.4xlarge"] }
+    },
 ];
 
 
@@ -64,7 +212,7 @@ local daskNodes = [
     metadata+: {
         name: "2i2c-aws-us",
         region: clusterRegion,
-        version: "1.29",
+        version: "1.30",
     },
     availabilityZones: masterAzs,
     iam: {
@@ -108,6 +256,9 @@ local daskNodes = [
                 "hub.jupyter.org/node-purpose": "core",
                 "k8s.dask.org/node-purpose": "core"
             },
+            tags+: {
+                "2i2c:node-purpose": "core"
+            },
         },
     ] + [
         ng + {
@@ -123,6 +274,9 @@ local daskNodes = [
                 "hub.jupyter.org/node-purpose": "user",
                 "k8s.dask.org/node-purpose": "scheduler"
             },
+            tags+: {
+                "2i2c:node-purpose": "user"
+            },
             taints+: {
                 "hub.jupyter.org_dedicated": "user:NoSchedule",
                 "hub.jupyter.org/dedicated": "user:NoSchedule"
@@ -145,6 +299,9 @@ local daskNodes = [
                 "k8s.dask.org_dedicated" : "worker:NoSchedule",
                 "k8s.dask.org/dedicated" : "worker:NoSchedule"
             },
+            tags+: {
+                "2i2c:node-purpose": "worker"
+            },
             instancesDistribution+: {
                 onDemandBaseCapacity: 0,
                 onDemandPercentageAboveBaseCapacity: 0,