fix: dependsOn bug fix for node rest use cases

.github/workflows/agentless-container.yaml

@@ -18,12 +18,6 @@
 # LICENSE END
 # 
 
-
-
-
-
-
-
 name: Build and push agentless container image
 
 # Configures this workflow to run every time a tag is created

.github/workflows/commit-linting.yaml

@@ -18,12 +18,6 @@
 # LICENSE END
 # 
 
-
-
-
-
-
-
 name: Commit Linting
 on: pull_request
 jobs:

.github/workflows/operator-ci.yaml

@@ -49,26 +49,26 @@ on:
 env:
   REGISTRY: ghcr.io
   IMAGE_NAME: ${{ github.repository }}
-  GO_VERSION: 1.23.7
+  GO_VERSION: 1.23.8
   PLATFORMS: linux/amd64,linux/arm64
 
 # There is a single job in this workflow. It's configured to run on the latest available version of Ubuntu.
 jobs:
-  unit-test:
+  ## move it all down to the tests job, should do all the tests then, and not double install stuff
+  # unit-test:
+  #   runs-on: ubuntu-latest
+  #   steps:
+  #     - uses: actions/checkout@v4
+  #     - name: Setup Go 1.23
+  #       uses: actions/setup-go@v5
+  #       with:
+  #         go-version: 1.23
+  #     - name: Unit tests
+  #       run: |
+  #         cd operator
+  #         make unit-tests
+  tests:
     runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@v4
-      - name: Setup Go 1.23
-        uses: actions/setup-go@v5
-        with:
-          go-version: 1.23
-      - name: Unit tests
-        run: |
-          cd operator
-          make unit-tests
-  k8s-tests:
-    runs-on: ubuntu-latest
-    needs: [unit-test] # Don't run the k8s tests if the unit tests fail
     steps:
       - uses: actions/checkout@v4
         with:
@@ -94,10 +94,10 @@ jobs:
         run: |
           cd operator
           GITHUB_TOKEN=${{ secrets.github_token }} make create-kind-cluster
-          make e2e-tests
+          make test
   build-and-push-operator:
     runs-on: ubuntu-latest
-    needs: [k8s-tests] # Don't run the build and push if the k8s tests fail
+    needs: [tests] # Don't run the build and push if the k8s tests fail
     # Sets the permissions granted to the `GITHUB_TOKEN` for the actions in this job.
     permissions:
       contents: read

.gitignore

@@ -1,4 +1,3 @@
 .cursorignore
 .pytest_cache
-.vscode
 .idea

.vscode/launch.json

@@ -0,0 +1,35 @@
+{
+    // Use IntelliSense to learn about possible attributes.
+    // Hover to view descriptions of existing attributes.
+    // For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387
+    "version": "0.2.0",
+    "configurations": [
+        {
+            "name": "Debug Manager",
+            "type": "go",
+            "request": "launch",
+            "mode": "debug",
+            "program": "${workspaceRoot}/operator/cmd/main.go",
+            "buildFlags": "--ldflags '-X github.com/NVIDIA/skyhook/internal/version.GIT_SHA=foobars -X github.com/NVIDIA/skyhook/internal/version.VERSION=v0.5.0'",
+            "env": {
+                "ENABLE_WEBHOOKS": "false",
+                "LOG_ENCODER": "console",
+                "REAPPLY_ON_REBOOT": "false",
+                // "AGENT_IMAGE": "ghcr.io/nvidia/skyhook/agent:latest", // the real agent image, for testing actual packages
+                "AGENT_IMAGE": "ghcr.io/nvidia/skyhook/agentless:6.2.0", // this is the mock image, if you need to test for real, use the real image.    
+            },
+            "args": [],
+            "showLog": true
+        },   
+        {
+            "name": "Test Current File",
+            "type": "go",
+            "request": "launch",
+            "mode": "test",
+            "program": "${file}",
+            "env": {},
+            "args": [],
+            "showLog": true
+        }   
+    ]
+}

agent/skyhook-agent/src/skyhook_agent/__about__.py

@@ -18,11 +18,4 @@
 # LICENSE END
 # 
 
-
-
-
-
-
-
-
 __version__ = "0.0.0"

agent/skyhook-agent/src/skyhook_agent/__init__.py

@@ -17,11 +17,3 @@
 #
 # LICENSE END
 # 
-
-
-
-
-
-
-
-

agent/skyhook-agent/src/skyhook_agent/controller.py

@@ -33,13 +33,26 @@
 import os
 import shutil
 import glob
+import signal
 
 from skyhook_agent.step import Step, UpgradeStep, Idempotence, Mode, CHECK_TO_APPLY
 from skyhook_agent import interrupts, config
 from typing import List
 
 import logging as logger
 
+# Global flag to track if we received SIGTERM
+received_sigterm = False
+
+def sigterm_handler(signum, frame):
+    """Handle SIGTERM by setting a global flag and logging the event"""
+    global received_sigterm
+    received_sigterm = True
+    logger.info("Received SIGTERM signal - initiating graceful shutdown")
+
+# Register the SIGTERM handler
+signal.signal(signal.SIGTERM, sigterm_handler)
+
 class SkyhookValidationError(Exception):
     pass
 
@@ -414,7 +427,11 @@ def remove_flags(step_data: dict[Mode, list[Step|UpgradeStep]], config_data: dic
         if os.path.exists(flag_file):  # Check if the file exists before trying to remove it
             os.remove(flag_file)
 
-def main(mode: Mode, root_mount: str, copy_dir: str, interrupt_data: None|str, always_run_step=False):
+def main(mode: Mode, root_mount: str, copy_dir: str, interrupt_data: None|str, always_run_step=False) -> bool:
+    '''
+    returns True if the there is a failure in the steps, otherwise returns False
+    '''
+
     if mode not in set(map(str, Mode)):
         logger.warning(f"This version of the Agent doesn't support the {mode} mode. Options are: {','.join(map(str, Mode))}.")
         return False
@@ -448,9 +465,19 @@ def main(mode: Mode, root_mount: str, copy_dir: str, interrupt_data: None|str, a
             if not os.path.exists(f"{root_mount}/{copy_dir}/configmaps/{f}"):
                 raise SkyhookValidationError(f"Expected config file {f} not found in configmaps directory.")
 
-    return agent_main(mode, root_mount, copy_dir, config_data, interrupt_data, always_run_step)
+    try:
+        return agent_main(mode, root_mount, copy_dir, config_data, interrupt_data, always_run_step)
+    except Exception as e:
+        if received_sigterm:
+            logger.info("Gracefully shutting down due to SIGTERM")
+            # Perform any cleanup if needed
+            return True
+        raise
 
-def agent_main(mode: Mode, root_mount: str, copy_dir: str, config_data: dict, interrupt_data: None|str, always_run_step=False):
+def agent_main(mode: Mode, root_mount: str, copy_dir: str, config_data: dict, interrupt_data: None|str, always_run_step=False) -> bool:
+    '''
+    returns True if the there is a failure in the steps, otherwise returns False
+    '''
 
     # Pull out step_data so it matches with existing code
     step_data = config_data["modes"]
@@ -464,6 +491,11 @@ def agent_main(mode: Mode, root_mount: str, copy_dir: str, config_data: dict, in
         logger.warning(f" There are no {mode} steps defined. This will be ran as a no-op.")
 
     for step in step_data.get(mode, []):
+        # Check for SIGTERM
+        if received_sigterm:
+            logger.info("SIGTERM received, stopping step execution")
+            return True
+
         # Make the flag file without the host path argument (first one). This is because in operator world
         # the host path is going to change every time the Skyhook Custom Resource changes so it would
         # look like a step hasn't been run when it fact it had.

agent/skyhook-agent/src/skyhook_agent/enums.py

@@ -18,13 +18,6 @@
 # LICENSE END
 # 
 
-
-
-
-
-
-
-
 from enum import Enum
 
 class SortableEnum(Enum):

agent/skyhook-agent/src/skyhook_agent/interrupts.py

@@ -19,12 +19,6 @@
 # 
 
 
-
-
-
-
-
-
 import string
 import base64
 import json

agent/skyhook-agent/src/skyhook_agent/step.py

@@ -19,12 +19,6 @@
 # 
 
 
-
-
-
-
-
-
 from typing import IO
 from enum import Enum
 import json

agent/skyhook-agent/tests/__init__.py

@@ -17,11 +17,3 @@
 #
 # LICENSE END
 # 
-
-
-
-
-
-
-
-

agent/skyhook-agent/tests/test_config.py

@@ -18,13 +18,6 @@
 # LICENSE END
 # 
 
-
-
-
-
-
-
-
 import unittest
 from tempfile import TemporaryDirectory
 

agent/skyhook-agent/tests/test_controller.py

@@ -18,6 +18,7 @@
 # LICENSE END
 # 
 
+
 import unittest
 import tempfile
 import os

agent/skyhook-agent/tests/test_enums.py

@@ -19,12 +19,6 @@
 # 
 
 
-
-
-
-
-
-
 import unittest
 
 from skyhook_agent.enums import SortableEnum, get_latest_schema

agent/skyhook-agent/tests/test_interrupts.py

@@ -19,12 +19,6 @@
 # 
 
 
-
-
-
-
-
-
 import unittest
 import base64
 import json

agent/skyhook-agent/tests/test_steps.py

@@ -19,12 +19,6 @@
 # 
 
 
-
-
-
-
-
-
 import unittest, os
 
 from tempfile import TemporaryDirectory

chart/templates/skyhook-crd.yaml

@@ -349,6 +349,10 @@ spec:
                         - name
                         type: object
                       type: array
+                    gracefulShutdown:
+                      description: GracefulShutdown is the graceful shutdown timeout
+                        for the package, if not set, uses k8s default
+                      type: string
                     image:
                       description: Image is the container image to run. Do not included
                         the tag, that is set in the version.

containers/agentless/Dockerfile

@@ -18,7 +18,6 @@
 # LICENSE END
 # 
 
-
 ARG BUSYBOX_TAG=1.36.1
 
 FROM busybox:${BUSYBOX_TAG}

containers/agentless/entrypoint.sh

@@ -20,6 +20,13 @@
 # LICENSE END
 # 
 
+# Handle SIGTERM gracefully
+cleanup() {
+    echo "Received SIGTERM signal, shutting down gracefully..."
+    sleep 3
+    exit 0
+}
+trap cleanup SIGTERM
 
 SLEEP_LEN=${SLEEP_LEN:-$(($RANDOM % 5 + 5))}
 

containers/ci.Dockerfile

@@ -18,6 +18,10 @@
 # LICENSE END
 # 
 
+## this container is not used in github CI, was from before before we open sourced this project.
+## should move to doing something like this in the github actions workflow to save time not installing all the deps all the time
+## but for now this is just for when we got to that
+
 ARG GO_VERSION
 
 FROM golang:${GO_VERSION}-bookworm as builder

demos/interrupt-wait-for-pod/non-workload.yaml

@@ -18,12 +18,6 @@
 # LICENSE END
 # 
 
-
-
-
-
-
-
 apiVersion: apps/v1
 kind: ReplicaSet
 metadata: