Merge pull request #2974 from consideRatio/pr/towards-single-dask-wor…

…ker-node-pool

gcp/aws, dask worker nodes: towards single r5.4xlarge/n2-highmem-16 dask worker node pool

eksctl/2i2c-aws-us.jsonnet

@@ -43,6 +43,10 @@ local daskNodes = [
     // *first* item in instanceDistribution.instanceTypes, to match
     // what we do with notebook nodes. Pods can request a particular
     // kind of node with a nodeSelector
+    //
+    // 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"] }},
 ];
 

eksctl/carbonplan.jsonnet

@@ -44,17 +44,18 @@ local notebookNodes = [
     },
 ];
 
-// Node definitions for dask worker nodes. Config here is merged
-// with our dask worker node definition, which uses spot instances.
-// A `node.kubernetes.io/instance-type label is set to the name of the
-// *first* item in instanceDistribution.instanceTypes, to match
-// what we do with notebook nodes. Pods can request a particular
-// kind of node with a nodeSelector
 local daskNodes = [
-    { instancesDistribution+: { instanceTypes: ["r5.large"] }},
-    { instancesDistribution+: { instanceTypes: ["r5.xlarge"] }},
-    { instancesDistribution+: { instanceTypes: ["r5.2xlarge"] }},
-    { instancesDistribution+: { instanceTypes: ["r5.8xlarge"] }},
+    // Node definitions for dask worker nodes. Config here is merged
+    // with our dask worker node definition, which uses spot instances.
+    // A `node.kubernetes.io/instance-type label is set to the name of the
+    // *first* item in instanceDistribution.instanceTypes, to match
+    // what we do with notebook nodes. Pods can request a particular
+    // kind of node with a nodeSelector
+    //
+    // 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"] }},
 ];
 
 {

eksctl/gridsst.jsonnet

@@ -44,6 +44,10 @@ local daskNodes = [
     // *first* item in instanceDistribution.instanceTypes, to match
     // what we do with notebook nodes. Pods can request a particular
     // kind of node with a nodeSelector
+    //
+    // 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"] }},
 ];
 

eksctl/jupyter-meets-the-earth.jsonnet

@@ -64,9 +64,11 @@ local daskNodes = [
     // *first* item in instanceDistribution.instanceTypes, to match
     // what we do with notebook nodes. Pods can request a particular
     // kind of node with a nodeSelector
-    { instancesDistribution+: { instanceTypes: ["m5.large"] }},
-    { instancesDistribution+: { instanceTypes: ["m5.4xlarge"] }},
-    { instancesDistribution+: { instanceTypes: ["m5.16xlarge"] }},
+    //
+    // 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"] }},
 ];
 
 

eksctl/nasa-cryo.jsonnet

@@ -37,6 +37,10 @@ local daskNodes = [
     // *first* item in instanceDistribution.instanceTypes, to match
     // what we do with notebook nodes. Pods can request a particular
     // kind of node with a nodeSelector
+    //
+    // 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"] }},
 ];
 

eksctl/nasa-ghg.jsonnet

@@ -36,6 +36,10 @@ local daskNodes = [
     // *first* item in instanceDistribution.instanceTypes, to match
     // what we do with notebook nodes. Pods can request a particular
     // kind of node with a nodeSelector
+    //
+    // 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"] }},
 ];
 

eksctl/nasa-veda.jsonnet

@@ -37,6 +37,10 @@ local daskNodes = [
     // *first* item in instanceDistribution.instanceTypes, to match
     // what we do with notebook nodes. Pods can request a particular
     // kind of node with a nodeSelector
+    //
+    // 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"] }},
 ];
 

eksctl/openscapes.jsonnet

@@ -41,9 +41,11 @@ local daskNodes = [
     // *first* item in instanceDistribution.instanceTypes, to match
     // what we do with notebook nodes. Pods can request a particular
     // kind of node with a nodeSelector
-    { instancesDistribution+: { instanceTypes: ["r5.xlarge"] }},
+    //
+    // 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"] }},
-    { instancesDistribution+: { instanceTypes: ["r5.16xlarge"] }},
 ];
 
 

eksctl/smithsonian.jsonnet

@@ -36,6 +36,10 @@ local daskNodes = [
     // *first* item in instanceDistribution.instanceTypes, to match
     // what we do with notebook nodes. Pods can request a particular
     // kind of node with a nodeSelector
+    //
+    // 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"] }},
 ];
 

eksctl/template.jsonnet

@@ -48,6 +48,10 @@ local daskNodes = [
     // *first* item in instanceDistribution.instanceTypes, to match
     // what we do with notebook nodes. Pods can request a particular
     // kind of node with a nodeSelector
+    //
+    // 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"] }},
 ];
 <% else %>

eksctl/victor.jsonnet

@@ -38,6 +38,10 @@ local daskNodes = [
     // *first* item in instanceDistribution.instanceTypes, to match
     // what we do with notebook nodes. Pods can request a particular
     // kind of node with a nodeSelector
+    //
+    // 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"] }},
 ];
 

terraform/gcp/projects/awi-ciroh.tfvars

@@ -48,6 +48,11 @@ notebook_nodes = {
   },
 }
 
+# Setup a single node pool for dask workers.
+#
+# A not yet fully established policy is being developed about using a single
+# node pool, see https://github.com/2i2c-org/infrastructure/issues/2687.
+#
 dask_nodes = {
   "medium" : {
     min : 0,

terraform/gcp/projects/cloudbank.tfvars

@@ -22,11 +22,16 @@ notebook_nodes = {
   },
 }
 
+# Setup a single node pool for dask workers.
+#
+# A not yet fully established policy is being developed about using a single
+# node pool, see https://github.com/2i2c-org/infrastructure/issues/2687.
+#
 dask_nodes = {
   "worker" : {
     min : 0,
     max : 100,
-    machine_type : "n1-highmem-4"
+    machine_type : "n2-highmem-16"
   },
 }
 

terraform/gcp/projects/daskhub-template.tfvars

@@ -67,8 +67,13 @@ notebook_nodes = {
   },
 }
 
+# Setup a single node pool for dask workers.
+#
+# A not yet fully established policy is being developed about using a single
+# node pool, see https://github.com/2i2c-org/infrastructure/issues/2687.
+#
 dask_nodes = {
-  "medium" : {
+  "worker" : {
     min : 0,
     max : 200,
     machine_type : "n2-highmem-16",

terraform/gcp/projects/leap.tfvars

@@ -97,6 +97,11 @@ notebook_nodes = {
   },
 }
 
+# Setup a single node pool for dask workers.
+#
+# A not yet fully established policy is being developed about using a single
+# node pool, see https://github.com/2i2c-org/infrastructure/issues/2687.
+#
 dask_nodes = {
   "medium" : {
     min : 0,

terraform/gcp/projects/linked-earth.tfvars

@@ -32,12 +32,17 @@ notebook_nodes = {
   },
 }
 
+# Setup a single node pool for dask workers.
+#
+# A not yet fully established policy is being developed about using a single
+# node pool, see https://github.com/2i2c-org/infrastructure/issues/2687.
+#
 dask_nodes = {
-  "medium" : {
+  "worker" : {
     min : 0,
     max : 100,
-    machine_type : "e2-highmem-16"
-  },
+    machine_type : "n2-highmem-16",
+  }
 }
 
 hub_cloud_permissions = {

terraform/gcp/projects/m2lines.tfvars

@@ -70,27 +70,17 @@ notebook_nodes = {
   }
 }
 
+# Setup a single node pool for dask workers.
+#
+# A not yet fully established policy is being developed about using a single
+# node pool, see https://github.com/2i2c-org/infrastructure/issues/2687.
+#
 dask_nodes = {
-  "small" : {
-    min : 0,
-    max : 100,
-    machine_type : "n1-standard-2",
-  },
-  "medium" : {
+  "worker" : {
     min : 0,
     max : 100,
-    machine_type : "n1-standard-4",
-  },
-  "large" : {
-    min : 0,
-    max : 100,
-    machine_type : "n1-standard-8",
-  },
-  "huge" : {
-    min : 0,
-    max : 100,
-    machine_type : "n1-standard-16",
-  },
+    machine_type : "n2-highmem-16",
+  }
 }
 
 hub_cloud_permissions = {

terraform/gcp/projects/meom-ige.tfvars

@@ -40,33 +40,17 @@ notebook_nodes = {
 
 }
 
+# Setup a single node pool for dask workers.
+#
+# A not yet fully established policy is being developed about using a single
+# node pool, see https://github.com/2i2c-org/infrastructure/issues/2687.
+#
 dask_nodes = {
-  "small" : {
-    min : 0,
-    max : 20,
-    machine_type : "n1-standard-2",
-  },
-  "medium" : {
+  "worker" : {
     min : 0,
-    max : 20,
-    machine_type : "n1-standard-8",
-  },
-  "large" : {
-    min : 0,
-    max : 20,
-    machine_type : "n1-standard-16",
-  },
-  "very-large" : {
-    min : 0,
-    max : 20,
-    machine_type : "n1-standard-32",
-  },
-  "huge" : {
-    min : 0,
-    max : 20,
-    machine_type : "n1-standard-64",
-  },
-
+    max : 100,
+    machine_type : "n2-highmem-16",
+  }
 }
 
 user_buckets = {

terraform/gcp/projects/pangeo-hubs.tfvars

@@ -1,3 +1,17 @@
+# SETTING UP TO WORK WITH THIS FILE:
+# -------------------------------------------------------------------------------
+#
+# The terraform state associated with this file is stored in a dedicated GCP
+# bucket, so in order to work with this file you need to do the following after
+# clearing a local .terraform folder.
+#
+# terraform init -backend-config backends/pangeo-backend.hcl
+# terraform workspace list
+# terraform workspace select <...>
+#
+# The GCP project having the bucket is https://console.cloud.google.com/?project=columbia
+#
+
 prefix                 = "pangeo-hubs"
 project_id             = "pangeo-integration-te-3eea"
 zone                   = "us-central1-b"
@@ -75,6 +89,24 @@ notebook_nodes = {
   },
 }
 
+# Setup a single node pool for dask workers.
+#
+# A not yet fully established policy is being developed about using a single
+# node pool, see https://github.com/2i2c-org/infrastructure/issues/2687.
+#
+# TODO: Transition to a single n2-highmem-16 worker node pool to be able to
+#       provide standardized worker pod config for all daskhubs.
+#
+#       Tracked in https://github.com/2i2c-org/infrastructure/issues/2687
+#
+#       The node pool to setup should look like this:
+#
+#       "worker" : {
+#         min : 0,
+#         max : 100,
+#         machine_type : "n2-highmem-16",
+#       },
+#
 dask_nodes = {
   "small" : {
     min : 0,

terraform/gcp/projects/pilot-hubs.tfvars

@@ -56,12 +56,17 @@ notebook_nodes = {
   }
 }
 
+# Setup a single node pool for dask workers.
+#
+# A not yet fully established policy is being developed about using a single
+# node pool, see https://github.com/2i2c-org/infrastructure/issues/2687.
+#
 dask_nodes = {
   "worker" : {
     min : 0,
     max : 100,
-    machine_type : "n1-highmem-4",
-  }
+    machine_type : "n2-highmem-16",
+  },
 }
 
 user_buckets = {}