add envFrom

Signed-off-by: Garrett Goon <[email protected]>

Author: garrett361

tools/pytorchjob-generator/chart/README.md

@@ -41,6 +41,7 @@ customize the Jobs generated by the tool.
 | Key | Type | Default | Description |
 |-----|------|---------|-------------|
 | environmentVariables | array | `nil` | List of variables/values to be defined for all the ranks. Values can be literals or references to Kuberetes secrets or configmaps.  See [values.yaml](values.yaml) for examples of supported syntaxes.  NOTE: The following standard [PyTorch Distributed environment variables](https://pytorch.org/docs/stable/distributed.html#environment-variable-initialization) are set automatically and can be referenced in the commands without being set manually: WORLD_SIZE, RANK, MASTER_ADDR, MASTER_PORT. |
+| envFrom | array | `nil` | List of ConfigMaps or Secrets specifying environment variables.  See [values.yaml](values.yaml) for examples of supported syntaxes.  NOTE: the environmentVariables field takes precedence over envFrom. mlbatch also performs some automatic checks on the environmentVariables passed by the user, such as checking that the user does not specify NCCL_TOPO_FILE when topologyFileConfigMap is also provided.  These checks are *not* performed on any environment variables inherited from envFrom. |
 | sshGitCloneConfig | object | `nil` | Private GitHub clone support. See [values.yaml](values.yaml) for additional instructions. |
 | setupCommands | array | no custom commands are executed | List of custom commands to be ran at the beginning of the execution. Use `setupCommand` to clone code, download data, and change directories. |
 | mainProgram | string | `nil` | Name of the PyTorch program to be executed by `torchrun`. Please provide your program name here and NOT in "setupCommands" as this helm template provides the necessary "torchrun" arguments for the parallel execution. WARNING: this program is relative to the current path set by change-of-directory commands in "setupCommands". If no value is provided; then only `setupCommands` are executed and torchrun is elided. |

tools/pytorchjob-generator/chart/templates/appwrapper.yaml

@@ -116,6 +116,10 @@ spec:
                                       {{- include "mlbatch.volumes" . | indent 38 }}
                                       containers:
                                           - name: pytorch
+                                            {{- if .Values.envFrom }}
+                                            envFrom:
+                                            {{- toYaml .Values.envFrom | nindent 46 }}
+                                            {{- end }}
                                             image: {{ required "Please specify a 'containerImage' in the user file" .Values.containerImage }}
                                             imagePullPolicy: {{ .Values.imagePullPolicy | default "IfNotPresent" }}
                                             {{- include "mlbatch.securityContext" . | indent 44 }}
@@ -139,6 +143,10 @@ spec:
                                       {{- include "mlbatch.volumes" . | indent 38 }}
                                       containers:
                                           - name: pytorch
+                                            {{- if .Values.envFrom }}
+                                            envFrom:
+                                            {{- toYaml .Values.envFrom | nindent 46 }}
+                                            {{- end }}
                                             image: {{ required "Please specify a 'containerImage' in the user file" .Values.containerImage }}
                                             imagePullPolicy: {{ .Values.imagePullPolicy | default "IfNotPresent" }}
                                             {{- include "mlbatch.securityContext" . | indent 44 }}

tools/pytorchjob-generator/chart/values.schema.json

@@ -42,6 +42,10 @@
             { "type": "null" },
             { "type": "array" }
         ]},
+        "envFrom": { "oneOf": [
+            { "type": "null" },
+            { "type": "array" }
+        ]},
         "sshGitCloneConfig": { "oneOf": [
             { "type": "null" },
             {

tools/pytorchjob-generator/chart/values.yaml

@@ -101,6 +101,23 @@ environmentVariables:
 #         name: configmap-name
 #         key: configmap-key
 
+
+# -- (array) List of ConfigMaps or Secrets specifying environment variables.  See
+# [values.yaml](values.yaml) for examples of supported syntaxes.
+#
+# NOTE: the environmentVariables field takes precedence over envFrom. mlbatch also performs some
+# automatic checks on the environmentVariables passed by the user, such as checking that the user
+# does not specify NCCL_TOPO_FILE when topologyFileConfigMap is also provided.  These checks are
+# *not* performed on any environment variables inherited from envFrom.
+# @section -- Workload Specification
+envFrom:
+#    - secretRef
+#      name: my-secrets
+#    - secretRef
+#      name: my-other-secrets
+#    - configMapRef
+#      name: my-config-map
+
 # Private GitHub clone support.
 #
 #    0) Create a secret and configMap to enable Private GitHub cloning as documented for your organization.