Add AutoDL Pro runner controller

2026-10-02 01:51:33 +08:00
commit 9bd8b7cc1e
6 changed files with 756 additions and 0 deletions
@@ -0,0 +1,28 @@
+name: Controller CI
+
+on:
+  push:
+    branches: [main]
+
+jobs:
+  test:
+    runs-on: ces-control
+    timeout-minutes: 10
+    env:
+      GITEA_JOB_TOKEN: ${{ secrets.GITEA_TOKEN }}
+    steps:
+      - name: Checkout controller source
+        run: |
+          git init .
+          git remote add origin "$GITHUB_SERVER_URL/$GITHUB_REPOSITORY.git"
+          git -c http.extraHeader="Authorization: token $GITEA_JOB_TOKEN" fetch --depth 1 origin "$GITHUB_SHA"
+          git checkout --detach FETCH_HEAD
+
+      - name: Verify controller
+        run: |
+          python3 -m py_compile controller.py
+          python3 -m unittest discover -s tests -v
+
+      - name: Confirm no paid action ran
+        run: |
+          echo "Controller CI only validates code; it never creates an AutoDL instance."
@@ -0,0 +1,57 @@
+name: Manage AutoDL GPU runner
+
+on:
+  workflow_dispatch:
+    inputs:
+      operation:
+        description: discover, provision, status, or shutdown
+        required: true
+        default: discover
+      image_uuid:
+        description: Required only for provision
+        required: false
+        default: ""
+      confirm_spend:
+        description: Type CREATE to provision a billed instance
+        required: false
+        default: ""
+
+jobs:
+  manage:
+    runs-on: ces-control
+    timeout-minutes: 180
+    env:
+      AUTODL_API_TOKEN: ${{ secrets.AUTODL_API_TOKEN }}
+      TRAINING_RUNNER_REGISTRATION_TOKEN: ${{ secrets.TRAINING_RUNNER_REGISTRATION_TOKEN }}
+      GITEA_JOB_TOKEN: ${{ secrets.GITEA_TOKEN }}
+      OPERATION: ${{ inputs.operation }}
+      IMAGE_UUID: ${{ inputs.image_uuid }}
+      CONFIRM_SPEND: ${{ inputs.confirm_spend }}
+    steps:
+      - name: Checkout controller source
+        run: |
+          git init .
+          git remote add origin "$GITHUB_SERVER_URL/$GITHUB_REPOSITORY.git"
+          git -c http.extraHeader="Authorization: token $GITEA_JOB_TOKEN" fetch --depth 1 origin "$GITHUB_SHA"
+          git checkout --detach FETCH_HEAD
+
+      - name: Validate requested operation
+        run: |
+          case "$OPERATION" in
+            discover|provision|status|shutdown) ;;
+            *) echo "Unsupported operation" >&2; exit 2 ;;
+          esac
+
+      - name: Manage AutoDL Pro instance
+        run: |
+          python3 controller.py "$OPERATION" \
+            --image-uuid "$IMAGE_UUID" \
+            --confirm-spend "$CONFIRM_SPEND"
+
+      - name: Controller host details
+        if: always()
+        run: |
+          echo "host=$(hostname)"
+          date -Iseconds
+          test -f /root/.local/state/ces-autodl-controller/state.json && \
+            python3 -c 'import json; p="/root/.local/state/ces-autodl-controller/state.json"; d=json.load(open(p)); d.pop("root_password", None); print(json.dumps(d, indent=2))' || true