Only deploy profilelist to staging first

yuvipanda · Mar 26, 2024 · 01b4211 · 01b4211
1 parent 00e476d
commit 01b4211
Show file tree

Hide file tree

Showing 3 changed files with 84 additions and 77 deletions.
diff --git a/config/clusters/awi-ciroh/common.values.yaml b/config/clusters/awi-ciroh/common.values.yaml
@@ -45,83 +45,6 @@ basehub:
             - jameshalgren
             - arpita0911patel
             - karnesh
-    singleuser:
-      profileList:
-        # The mem-guarantees are here so k8s doesn't schedule other pods
-        # on these nodes. They need to be just under total allocatable
-        # RAM on a node, not total node capacity. Values calculated using
-        # https://learnk8s.io/kubernetes-instance-calculator
-        #
-        # FIXME: These are changed to a temporary node sharing setup based on
-        #        the legacy choices to help us pre-warm capacity independent on
-        #        the choices users end up making and avoiding running into
-        #        persistent disk quotas.
-        #
-        #        Change PR: https://github.com/2i2c-org/infrastructure/pull/2539
-        #        Related event: https://github.com/2i2c-org/infrastructure/issues/2520
-        #
-        #        This is an interim setup, trying to balance the experience of
-        #        the previous 1:1 user:node setup with a node sharing setup. It
-        #        is not meant to be retained long term!
-        #
-        #        -[ ] Make this cluster have a node sharing setup like in the
-        #             basehub/daskhub template.
-        #
-        - display_name: "Small"
-          description: 5GB RAM, 2 CPUs
-          default: true
-          profile_options: &profile_options
-            image:
-              display_name: Image
-              choices:
-                old:
-                  display_name: Original Pangeo Notebook base image 2023.01.13
-                  slug: "tensorflow"
-                  kubespawner_override:
-                    image: "quay.io/2i2c/awi-ciroh-image:f7222fce8b16 "
-                new:
-                  display_name: New Pangeo Notebook base image 2023.09.11
-                  default: true
-                  slug: "pytorch"
-                  kubespawner_override:
-                    image: "quay.io/2i2c/awi-ciroh-image:88ea9a74a66e"
-          kubespawner_override:
-            mem_limit: 7G
-            mem_guarantee: 5G
-            cpu_limit: 2
-            cpu_guarantee: 0.938
-            node_selector:
-              node.kubernetes.io/instance-type: n2-highmem-16
-        - display_name: Medium
-          description: 11GB RAM, 4 CPUs
-          profile_options: *profile_options
-          kubespawner_override:
-            mem_limit: 15G
-            mem_guarantee: 11G
-            cpu_limit: 4
-            cpu_guarantee: 1.875
-            node_selector:
-              node.kubernetes.io/instance-type: n2-highmem-16
-        - display_name: Large
-          description: 24GB RAM, 8 CPUs
-          profile_options: *profile_options
-          kubespawner_override:
-            mem_limit: 30G
-            mem_guarantee: 24G
-            cpu_limit: 8
-            cpu_guarantee: 3.75
-            node_selector:
-              node.kubernetes.io/instance-type: n2-highmem-16
-        - display_name: Huge
-          description: 52GB RAM, 16 CPUs
-          profile_options: *profile_options
-          kubespawner_override:
-            mem_limit: 60G
-            mem_guarantee: 52G
-            cpu_limit: 16
-            cpu_guarantee: 7.5
-            node_selector:
-              node.kubernetes.io/instance-type: n2-highmem-16
 dask-gateway:
   gateway:
     backend:

diff --git a/config/clusters/awi-ciroh/prod.values.yaml b/config/clusters/awi-ciroh/prod.values.yaml
@@ -9,6 +9,14 @@ basehub:
         - hosts: [ciroh.awi.2i2c.cloud]
           secretName: https-auto-tls
     singleuser:
+      image:
+        # Image build repo: https://github.com/2i2c-org/awi-ciroh-image
+        #
+        # NOTE: The configurator is used in this cluster, so this is stale
+        #       configuration.
+        #
+        name: "quay.io/2i2c/awi-ciroh-image"
+        tag: "584293e50d4c"
       extraEnv:
         SCRATCH_BUCKET: gs://awi-ciroh-scratch/$(JUPYTERHUB_USER)
         PANGEO_SCRATCH: gs://awi-ciroh-scratch/$(JUPYTERHUB_USER)

diff --git a/config/clusters/awi-ciroh/staging.values.yaml b/config/clusters/awi-ciroh/staging.values.yaml
@@ -9,6 +9,82 @@ basehub:
         - hosts: [staging.ciroh.awi.2i2c.cloud]
           secretName: https-auto-tls
     singleuser:
+      profileList:
+        # The mem-guarantees are here so k8s doesn't schedule other pods
+        # on these nodes. They need to be just under total allocatable
+        # RAM on a node, not total node capacity. Values calculated using
+        # https://learnk8s.io/kubernetes-instance-calculator
+        #
+        # FIXME: These are changed to a temporary node sharing setup based on
+        #        the legacy choices to help us pre-warm capacity independent on
+        #        the choices users end up making and avoiding running into
+        #        persistent disk quotas.
+        #
+        #        Change PR: https://github.com/2i2c-org/infrastructure/pull/2539
+        #        Related event: https://github.com/2i2c-org/infrastructure/issues/2520
+        #
+        #        This is an interim setup, trying to balance the experience of
+        #        the previous 1:1 user:node setup with a node sharing setup. It
+        #        is not meant to be retained long term!
+        #
+        #        -[ ] Make this cluster have a node sharing setup like in the
+        #             basehub/daskhub template.
+        #
+        - display_name: "Small"
+          description: 5GB RAM, 2 CPUs
+          default: true
+          profile_options: &profile_options
+            image:
+              display_name: Image
+              choices:
+                old:
+                  display_name: Original Pangeo Notebook base image 2023.01.13
+                  slug: "tensorflow"
+                  kubespawner_override:
+                    image: "quay.io/2i2c/awi-ciroh-image:f7222fce8b16 "
+                new:
+                  display_name: New Pangeo Notebook base image 2023.09.11
+                  default: true
+                  slug: "pytorch"
+                  kubespawner_override:
+                    image: "quay.io/2i2c/awi-ciroh-image:88ea9a74a66e"
+          kubespawner_override:
+            mem_limit: 7G
+            mem_guarantee: 5G
+            cpu_limit: 2
+            cpu_guarantee: 0.938
+            node_selector:
+              node.kubernetes.io/instance-type: n2-highmem-16
+        - display_name: Medium
+          description: 11GB RAM, 4 CPUs
+          profile_options: *profile_options
+          kubespawner_override:
+            mem_limit: 15G
+            mem_guarantee: 11G
+            cpu_limit: 4
+            cpu_guarantee: 1.875
+            node_selector:
+              node.kubernetes.io/instance-type: n2-highmem-16
+        - display_name: Large
+          description: 24GB RAM, 8 CPUs
+          profile_options: *profile_options
+          kubespawner_override:
+            mem_limit: 30G
+            mem_guarantee: 24G
+            cpu_limit: 8
+            cpu_guarantee: 3.75
+            node_selector:
+              node.kubernetes.io/instance-type: n2-highmem-16
+        - display_name: Huge
+          description: 52GB RAM, 16 CPUs
+          profile_options: *profile_options
+          kubespawner_override:
+            mem_limit: 60G
+            mem_guarantee: 52G
+            cpu_limit: 16
+            cpu_guarantee: 7.5
+            node_selector:
+              node.kubernetes.io/instance-type: n2-highmem-16
       extraEnv:
         SCRATCH_BUCKET: gs://awi-ciroh-scratch-staging/$(JUPYTERHUB_USER)
         PANGEO_SCRATCH: gs://awi-ciroh-scratch-staging/$(JUPYTERHUB_USER)