Update default container security context (#2265)

* Update default container security context

Signed-off-by: Yi Chen <[email protected]>

* Push user and group directives into Dockerfile

Signed-off-by: Yi Chen <[email protected]>

* Add allowPrivilegeEscalation to container security context

Signed-off-by: Yi Chen <[email protected]>

* fix: fsGroup should be moved to pod security context

Signed-off-by: Yi Chen <[email protected]>

---------

Signed-off-by: Yi Chen <[email protected]>

Author: ChenYi015

Dockerfile

@@ -26,7 +26,9 @@ RUN --mount=type=cache,target=/go/pkg/mod/ \
     go mod download
 
 COPY . .
+
 ENV GOCACHE=/root/.cache/go-build
+
 ARG TARGETARCH
 
 RUN --mount=type=cache,target=/go/pkg/mod/ \
@@ -35,6 +37,10 @@ RUN --mount=type=cache,target=/go/pkg/mod/ \
 
 FROM ${SPARK_IMAGE}
 
+ARG SPARK_UID=185
+
+ARG SPARK_GID=185
+
 USER root
 
 RUN apt-get update \
@@ -45,7 +51,7 @@ RUN mkdir -p /etc/k8s-webhook-server/serving-certs /home/spark && \
     chmod -R g+rw /etc/k8s-webhook-server/serving-certs && \
     chown -R spark /etc/k8s-webhook-server/serving-certs /home/spark
 
-USER spark
+USER ${SPARK_UID}:${SPARK_GID}
 
 COPY --from=builder /workspace/bin/spark-operator /usr/bin/spark-operator
 

charts/spark-operator-chart/README.md

@@ -106,13 +106,13 @@ See [helm uninstall](https://helm.sh/docs/helm/helm_uninstall) for command docum
 | controller.affinity | object | `{}` | Affinity for controller pods. |
 | controller.tolerations | list | `[]` | List of node taints to tolerate for controller pods. |
 | controller.priorityClassName | string | `""` | Priority class for controller pods. |
-| controller.podSecurityContext | object | `{}` | Security context for controller pods. |
+| controller.podSecurityContext | object | `{"fsGroup":185}` | Security context for controller pods. |
 | controller.topologySpreadConstraints | list | `[]` | Topology spread constraints rely on node labels to identify the topology domain(s) that each Node is in. Ref: [Pod Topology Spread Constraints](https://kubernetes.io/docs/concepts/workloads/pods/pod-topology-spread-constraints/). The labelSelector field in topology spread constraint will be set to the selector labels for controller pods if not specified. |
 | controller.env | list | `[]` | Environment variables for controller containers. |
 | controller.envFrom | list | `[]` | Environment variable sources for controller containers. |
 | controller.volumeMounts | list | `[]` | Volume mounts for controller containers. |
 | controller.resources | object | `{}` | Pod resource requests and limits for controller containers. Note, that each job submission will spawn a JVM within the controller pods using "/usr/local/openjdk-11/bin/java -Xmx128m". Kubernetes may kill these Java processes at will to enforce resource limits. When that happens, you will see the following error: 'failed to run spark-submit for SparkApplication [...]: signal: killed' - when this happens, you may want to increase memory limits. |
-| controller.securityContext | object | `{}` | Security context for controller containers. |
+| controller.securityContext | object | `{"allowPrivilegeEscalation":false,"capabilities":{"drop":["ALL"]},"privileged":false,"runAsNonRoot":true}` | Security context for controller containers. |
 | controller.sidecars | list | `[]` | Sidecar containers for controller pods. |
 | controller.podDisruptionBudget.enable | bool | `false` | Specifies whether to create pod disruption budget for controller. Ref: [Specifying a Disruption Budget for your Application](https://kubernetes.io/docs/tasks/run-application/configure-pdb/) |
 | controller.podDisruptionBudget.minAvailable | int | `1` | The number of pods that must be available. Require `controller.replicas` to be greater than 1 |
@@ -144,13 +144,13 @@ See [helm uninstall](https://helm.sh/docs/helm/helm_uninstall) for command docum
 | webhook.affinity | object | `{}` | Affinity for webhook pods. |
 | webhook.tolerations | list | `[]` | List of node taints to tolerate for webhook pods. |
 | webhook.priorityClassName | string | `""` | Priority class for webhook pods. |
-| webhook.podSecurityContext | object | `{}` | Security context for webhook pods. |
+| webhook.podSecurityContext | object | `{"fsGroup":185}` | Security context for webhook pods. |
 | webhook.topologySpreadConstraints | list | `[]` | Topology spread constraints rely on node labels to identify the topology domain(s) that each Node is in. Ref: [Pod Topology Spread Constraints](https://kubernetes.io/docs/concepts/workloads/pods/pod-topology-spread-constraints/). The labelSelector field in topology spread constraint will be set to the selector labels for webhook pods if not specified. |
 | webhook.env | list | `[]` | Environment variables for webhook containers. |
 | webhook.envFrom | list | `[]` | Environment variable sources for webhook containers. |
 | webhook.volumeMounts | list | `[]` | Volume mounts for webhook containers. |
 | webhook.resources | object | `{}` | Pod resource requests and limits for webhook pods. |
-| webhook.securityContext | object | `{}` | Security context for webhook containers. |
+| webhook.securityContext | object | `{"allowPrivilegeEscalation":false,"capabilities":{"drop":["ALL"]},"privileged":false,"runAsNonRoot":true}` | Security context for webhook containers. |
 | webhook.podDisruptionBudget.enable | bool | `false` | Specifies whether to create pod disruption budget for webhook. Ref: [Specifying a Disruption Budget for your Application](https://kubernetes.io/docs/tasks/run-application/configure-pdb/) |
 | webhook.podDisruptionBudget.minAvailable | int | `1` | The number of pods that must be available. Require `webhook.replicas` to be greater than 1 |
 | spark.jobNamespaces | list | `["default"]` | List of namespaces where to run spark jobs. If empty string is included, all namespaces will be allowed. Make sure the namespaces have already existed. |

charts/spark-operator-chart/values.yaml

@@ -120,10 +120,8 @@ controller:
   priorityClassName: ""
 
   # -- Security context for controller pods.
-  podSecurityContext: {}
-    # runAsUser: 1000
-    # runAsGroup: 2000
-    # fsGroup: 3000
+  podSecurityContext:
+    fsGroup: 185
 
   # -- Topology spread constraints rely on node labels to identify the topology domain(s) that each Node is in.
   # Ref: [Pod Topology Spread Constraints](https://kubernetes.io/docs/concepts/workloads/pods/pod-topology-spread-constraints/).
@@ -158,10 +156,13 @@ controller:
     #   memory: 300Mi
 
   # -- Security context for controller containers.
-  securityContext: {}
-    # runAsUser: 1000
-    # runAsGroup: 2000
-    # fsGroup: 3000
+  securityContext:
+    privileged: false
+    allowPrivilegeEscalation: false
+    runAsNonRoot: true
+    capabilities:
+      drop:
+      - ALL
 
   # -- Sidecar containers for controller pods.
   sidecars: []
@@ -266,10 +267,8 @@ webhook:
   priorityClassName: ""
 
   # -- Security context for webhook pods.
-  podSecurityContext: {}
-    # runAsUser: 1000
-    # runAsGroup: 2000
-    # fsGroup: 3000
+  podSecurityContext:
+    fsGroup: 185
 
   # -- Topology spread constraints rely on node labels to identify the topology domain(s) that each Node is in.
   # Ref: [Pod Topology Spread Constraints](https://kubernetes.io/docs/concepts/workloads/pods/pod-topology-spread-constraints/).
@@ -301,10 +300,13 @@ webhook:
     #   memory: 300Mi
 
   # -- Security context for webhook containers.
-  securityContext: {}
-    # runAsUser: 1000
-    # runAsGroup: 2000
-    # fsGroup: 3000
+  securityContext:
+    privileged: false
+    allowPrivilegeEscalation: false
+    runAsNonRoot: true
+    capabilities:
+      drop:
+      - ALL
 
   # Pod disruption budget for webhook to avoid service degradation.
   podDisruptionBudget: