add env to triton

charts/hypha/Chart.lock

@@ -7,9 +7,9 @@ dependencies:
   version: 3.3.0
 - name: tritoninferenceserver-hypha
   repository: file://../tritoninferenceserver-hypha
-  version: 1.0.5
+  version: 1.0.7
 - name: redis
   repository: https://charts.bitnami.com/bitnami
   version: 17.11.6
-digest: sha256:f513dba8f9978281f7d8e496c3341cdcabb07389f08fe168e52869b186acc852
-generated: "2023-09-19T10:41:56.049018409+01:00"
+digest: sha256:71a65bfbce59f1f2a3284ea4609953afd68a07856c039442322110994c89c85e
+generated: "2023-10-12T16:33:20.113826467+01:00"

charts/hypha/Chart.yaml

@@ -13,7 +13,7 @@ type: application
 # This is the chart version. This version number should be incremented each time you make changes
 # to the chart and its templates, including the app version.
 # Versions are expected to follow Semantic Versioning (https://semver.org/)
-version: 0.16.1
+version: 0.16.3
 # This is the version number of the application being deployed. This version number should be
 # incremented each time you make changes to the application. Versions are not expected to
 # follow Semantic Versioning. They should reflect the version the application is using.
@@ -43,7 +43,7 @@ dependencies:
   #   condition: ingress-nginx.enabled
   - name: tritoninferenceserver-hypha
     repository: "file://../tritoninferenceserver-hypha"
-    version: 1.0.5
+    version: 1.0.7
     condition: tritoninferenceserver-hypha.enabled
   - name: redis
     repository: "https://charts.bitnami.com/bitnami"

charts/hypha/values.yaml

@@ -101,6 +101,19 @@ tritoninferenceserver-hypha:
     numGpus: 0
     # s3Endpoint: https://uk1s3.embassy.ebi.ac.uk
     # s3Bucket: s3://model-repository/
+    env:
+      - name: TF_FORCE_GPU_ALLOW_GROWTH
+        value: "true"
+      - name: AWS_ACCESS_KEY_ID
+        value: minio
+      - name: AWS_SECRET_ACCESS_KEY
+        value: minio123
+      - name: AWS_REGION
+        value: us-east-1
+      - name: S3_ENDPOINT
+        value: "http://minio:9000"
+      - name: S3_USE_HTTPS
+        value: "0"
   ingress:
     enabled: false
   # resources:
@@ -114,13 +127,23 @@ tritoninferenceserver-hypha:
   #     cpu: "4"
   resources:
     requests:
-      nvidia.com/gpu: 0
+      nvidia.com/gpu: 1
       memory: "1Gi"
       cpu: "0.1"
     limits:
-      nvidia.com/gpu: 0
+      nvidia.com/gpu: 1
       memory: "1Gi"
       cpu: "0.1"
 
   service:
     type: ClusterIP
+
+  autoscaling:
+    enabled: false
+    minReplicas: 1
+    maxReplicas: 100
+    gpuUtilization: 50
+    gpuMemoryUtilization: 50
+
+  nodeSelector:
+    cloud.google.com/gke-accelerator: "nvidia-tesla-t4"

charts/tritoninferenceserver-hypha/Chart.yaml

@@ -15,7 +15,7 @@ type: application
 # This is the chart version. This version number should be incremented each time you make changes
 # to the chart and its templates, including the app version.
 # Versions are expected to follow Semantic Versioning (https://semver.org/)
-version: 1.0.5
+version: 1.0.7
 
 # This is the version number of the application being deployed. This version number should be
 # incremented each time you make changes to the application. Versions are not expected to

charts/tritoninferenceserver-hypha/templates/deployment.yaml

@@ -41,10 +41,9 @@ spec:
             - mountPath:  {{ .Values.image.modelSnapshotsPath }}
               name: model-snapshots
           env:
-          - name: TF_FORCE_GPU_ALLOW_GROWTH
-            value: "true"
-          - name: MODEL_SNAPSHOTS_DIRECTORY
-            value: "{{ .Values.image.modelSnapshotsPath }}"
+            - name: TF_FORCE_GPU_ALLOW_GROWTH
+              value: "true"
+{{ toYaml .Values.image.env | indent 12 }}
           imagePullPolicy: IfNotPresent
         #   TODO Fix probe checks
         #   livenessProbe:

charts/tritoninferenceserver-hypha/values.yaml

@@ -8,6 +8,9 @@ image:
   numGpus: 0
   s3Endpoint: https://uk1s3.embassy.ebi.ac.uk
   s3Bucket: s3://model-repository/
+  env:
+    - name: TF_FORCE_GPU_ALLOW_GROWTH
+      value: "true"
 ingress:
   enabled: false
   className: nginx