--- controls: version: 3.10 id: 1 text: "Securing the OpenShift Master" type: "master" groups: - id: 1 text: "Protecting the API Server" checks: - id: 1.1 text: "Maintain default behavior for anonymous access" type: "skip" scored: true - id: 1.2 text: "Verify that the basic-auth-file method is not enabled" audit: "grep -A2 basic-auth-file /etc/origin/master/master-config.yaml" tests: test_items: - flag: "--basic-auth-file" compare: op: eq value: "" set: false remediation: | Edit the kubernetes master config file /etc/origin/master/master-config.yaml and remove the basic-auth-file entry. kubernetesMasterConfig:  apiServerArguments:    basic-auth-file:    - /path/to/any/file scored: true - id: 1.3 text: "Insecure Tokens" type: "skip" scored: true - id: 1.4 text: "Secure communications between the API server and master nodes" audit: "grep -A4 kubeletClientInfo /etc/origin/master/master-config.yaml" tests: bin_op: and test_items: - flag: "kubeletClientInfo:" compare: op: eq value: "kubeletClientInfo:" set: true - flag: "ca" compare: op: has value: "ca-bundle.crt" set: true - flag: "certFile" compare: op: has value: "master.kubelet-client.crt" set: true - flag: "keyFile" compare: op: has value: "master.kubelet-client.key" set: true - flag: "port: 10250" compare: op: eq value: "port: 10250" set: true remediation: | Edit the kubernetes master config file /etc/origin/master/master-config.yaml and change it to match the below. kubeletClientInfo:  ca: ca-bundle.crt  certFile: master.kubelet-client.crt  keyFile: master.kubelet-client.key  port: 10250 scored: true - id: 1.5 text: "Prevent insecure bindings" audit: "grep -A2 insecure-bind-address /etc/origin/master/master-config.yaml" tests: test_items: - flag: "insecure-bind-address" set: false remediation: | Edit the kubernetes master config file /etc/origin/master/master-config.yaml and remove the insecure-bind-address entry. kubernetesMasterConfig:  apiServerArguments:    insecure-bind-address:    - 127.0.0.1 scored: true - id: 1.6 text: "Prevent insecure port access" audit: "grep -A2 insecure-port /etc/origin/master/master-config.yaml" tests: test_items: - flag: "insecure-port" set: false remediation: | Edit the kubernetes master config file /etc/origin/master/master-config.yaml and remove the insecure-port entry. kubernetesMasterConfig:  apiServerArguments:   insecure-port:  - 0 scored: true - id: 1.7 text: "Use Secure Ports for API Server Traffic" audit: "grep -A2 secure-port /etc/origin/master/master-config.yaml" tests: bin_op: or test_items: - flag: "secure-port" set: false - flag: "secure-port" compare: op: nothave value: "0" set: true remediation: | Edit the kubernetes master config file /etc/origin/master/master-config.yaml and either remove the secure-port parameter or set it to a different (non-zero) desired port. kubernetesMasterConfig:  apiServerArguments:   secure-port:  - 8443 scored: true - id: 1.8 text: "Do not expose API server profiling data" type: "skip" scored: true - id: 1.9 text: "Verify repair-malformed-updates argument for API compatibility" audit: "grep -A2 repair-malformed-updates /etc/origin/master/master-config.yaml" tests: bin_op: or test_items: - flag: "repair-malformed-updates" set: false - flag: "repair-malformed-updates" compare: op: has value: "true" set: true remediation: | Edit the kubernetes master config file /etc/origin/master/master-config.yaml and remove the repair-malformed-updates entry or set repair-malformed-updates=true. scored: true - id: 1.10 text: "Verify that the AlwaysAdmit admission controller is disabled" audit: "grep -A4 AlwaysAdmit /etc/origin/master/master-config.yaml" tests: test_items: - flag: "AlwaysAdmit" set: false remediation: | Edit the kubernetes master config file /etc/origin/master/master-config.yaml and remove the entry below. AlwaysAdmit: configuration: kind: DefaultAdmissionConfig apiVersion: v1 disable: false scored: true - id: 1.11 text: "Manage the AlwaysPullImages admission controller" audit: "grep -A4 AlwaysPullImages /etc/origin/master/master-config.yaml" tests: test_items: - flag: "disable" compare: op: has value: "false" set: true remediation: | Edit the kubernetes master config file /etc/origin/master/master-config.yaml and add the entry below. admissionConfig: pluginConfig: AlwaysPullImages: configuration: kind: DefaultAdmissionConfig apiVersion: v1 disable: false scored: true - id: 1.12 text: "Use Security Context Constraints instead of DenyEscalatingExec admission" type: "skip" scored: true - id: 1.13 text: "Use Security Context Constraints instead of the SecurityContextDeny admission controller" type: "skip" scored: true - id: 1.14 text: "Manage the NamespaceLifecycle admission controller" audit: "grep -A4 NamespaceLifecycle /etc/origin/master/master-config.yaml" tests: test_items: - flag: "NamespaceLifecycle" set: false remediation: | Edit the kubernetes master config file /etc/origin/master/master-config.yaml and remove the following entry. NamespaceLifecycle: configuration: kind: DefaultAdmissionConfig apiVersion: v1 disable: true scored: true - id: 1.15 text: "Configure API server auditing - audit log file path" audit: "grep -A5 auditConfig /etc/origin/master/master-config.yaml" tests: test_items: - flag: "enabled" compare: op: has value: "true" set: true remediation: | Edit the Openshift master config file /etc/origin/master/master-config.yaml, update the following entry and restart the API server. auditConfig: auditFilePath: ""/etc/origin/master/audit-ocp.log"" enabled: true maximumFileRetentionDays: 30 maximumFileSizeMegabytes: 10 maximumRetainedFiles: 10 Make the same changes in the inventory/ansible variables so the changes are not lost when an upgrade occurs. scored: true - id: 1.16 text: "Configure API server auditing - audit log retention" audit: "grep -A5 auditConfig /etc/origin/master/master-config.yaml" tests: test_items: - flag: "maximumFileRetentionDays: 30" compare: op: has value: "maximumFileRetentionDays" set: true remediation: | Edit the Openshift master config file /etc/origin/master/master-config.yaml, update the maximumFileRetentionDays entry and restart the API server. auditConfig: auditFilePath: ""/etc/origin/master/audit-ocp.log"" enabled: true maximumFileRetentionDays: 30 maximumFileSizeMegabytes: 10 maximumRetainedFiles: 10 Make the same changes in the inventory/ansible variables so the changes are not lost when an upgrade occurs. scored: true - id: 1.17 text: "Configure API server auditing - audit log backup retention" audit: "grep -A5 auditConfig /etc/origin/master/master-config.yaml" tests: test_items: - flag: "maximumRetainedFiles: 10" compare: op: has value: "maximumRetainedFiles" set: true remediation: | Edit the Openshift master config file /etc/origin/master/master-config.yaml, update the maximumRetainedFiles entry, set enabled to true and restart the API server. auditConfig: auditFilePath: ""/etc/origin/master/audit-ocp.log"" enabled: true maximumFileRetentionDays: 30 maximumFileSizeMegabytes: 10 maximumRetainedFiles: 10 Make the same changes in the inventory/ansible variables so the changes are not lost when an upgrade occurs. scored: true - id: 1.18 text: "Configure audit log file size" audit: "grep -A5 auditConfig /etc/origin/master/master-config.yaml" tests: test_items: - flag: "maximumFileSizeMegabytes: 30" compare: op: has value: "maximumFileSizeMegabytes" set: true remediation: | Edit the Openshift master config file /etc/origin/master/master-config.yaml, update the maximumFileSizeMegabytes entry, set enabled to true and restart the API server. auditConfig: auditFilePath: ""/etc/origin/master/audit-ocp.log"" enabled: true maximumFileRetentionDays: 30 maximumFileSizeMegabytes: 10 maximumRetainedFiles: 10 Make the same changes in the inventory/ansible variables so the changes are not lost when an upgrade occurs. scored: true - id: 1.19 text: "Verify that authorization-mode is not set to AlwaysAllow" audit: "grep -A1 authorization-mode /etc/origin/master/master-config.yaml" tests: test_items: - flag: "authorization-mode" set: false remediation: | Edit the Openshift master config file /etc/origin/master/master-config.yaml and remove the authorization-mode entry. kubernetesMasterConfig:  apiServerArguments:    authorization-mode:    - AllowAll scored: true - id: 1.20 text: "Verify that the token-auth-file flag is not set" audit: "grep token-auth-file /etc/origin/master/master-config.yaml" tests: test_items: - flag: "token-auth-file" set: false remediation: | Edit the Openshift master config file /etc/origin/master/master-config.yaml and remove the token-auth-file entry under apiserverArguments section. kubernetesMasterConfig:  apiServerArguments:    token-auth-file:    - /path/to/file scored: true - id: 1.21 text: "Verify the API server certificate authority" audit: "grep -A1 kubelet-certificate-authority /etc/origin/master/master-config.yaml" tests: test_items: - flag: "kubelet-certificate-authority" set: false remediation: | Edit the Openshift master config file /etc/origin/master/master-config.yaml and remove the following configuration under apiserverArguments section. kubernetesMasterConfig:  apiServerArguments:    kubelet-certificat-authority:    - /path/to/ca scored: true - id: 1.22 text: "Verify the API server client certificate and client key" audit: "grep -A4 kubeletClientInfo /etc/origin/master/master-config.yaml" tests: bin_op: and test_items: - flag: "keyFile" compare: op: has value: "master.kubelet-client.key" set: true - flag: "certFile" compare: op: has value: "master.kubelet-client.crt" set: true remediation: | Edit the Openshift master config file /etc/origin/master/master-config.yaml and add the following configuration under kubeletClientInfo kubeletClientInfo:  ca: ca-bundle.crt  certFile: master.kubelet-client.crt  keyFile: master.kubelet-client.key port: 10250 scored: true - id: 1.23 text: "Verify that the service account lookup flag is not set" type: "skip" scored: true - id: 1.24 text: "Verify the PodSecurityPolicy is disabled to ensure use of SecurityContextConstraints" type: "skip" scored: true - id: 1.25 text: "Verify that the service account key file argument is not set" audit: "grep -A9 serviceAccountConfig /etc/origin/master/master-config.yaml" tests: bin_op: and test_items: - flag: "privateKeyFile" compare: op: has value: "serviceaccounts.private.key" set: true - flag: "serviceaccounts.public.key" compare: op: has value: "serviceaccounts.public.key" set: true remediation: | OpenShift API server does not use the service-account-key-file argument. Even if value is set in master-config.yaml, it will not be used to verify service account tokens, as it is in upstream Kubernetes. The ServiceAccount token authenticator is configured with serviceAccountConfig.publicKeyFiles in the master-config.yaml. OpenShift does not reuse the apiserver TLS key. Edit the Openshift master config file /etc/origin/master/master-config.yaml and set the privateKeyFile and publicKeyFile configuration under serviceAccountConfig. serviceAccountConfig:  limitSecretReferences: false  managedNames: - default  - builder  - deployer  masterCA: ca-bundle.crt   privateKeyFile: serviceaccounts.private.key  publicKeyFiles:  - serviceaccounts.public.key Verify that privateKeyFile and publicKeyFile exist and set. scored: true - id: 1.26 text: "Verify the certificate and key used for communication with etcd" audit: "grep -A3 etcdClientInfo /etc/origin/master/master-config.yaml" tests: bin_op: and test_items: - flag: "certFile" compare: op: has value: "master.etcd-client.crt" set: true - flag: "keyFile" compare: op: has value: "master.etcd-client.key" set: true remediation: | Edit the Openshift master config file /etc/origin/master/master-config.yaml and set keyFile and certFile under etcdClientInfo like below. etcdClientInfo:  ca: master.etcd-ca.crt certFile: master.etcd-client.crt keyFile: master.etcd-client.key scored: true - id: 1.27 text: "Verify that the ServiceAccount admission controller is enabled" audit: "grep -A4 ServiceAccount /etc/origin/master/master-config.yaml" tests: bin_op: or test_items: - flag: "ServiceAccount" set: false - flag: "disable" compare: op: has value: "false" set: true remediation: | Edit the Openshift master config file /etc/origin/master/master-config.yaml and enable ServiceAccount admission control policy. ServiceAccount: configuration: kind: DefaultAdmissionConfig apiVersion: v1 disable: false scored: true - id: 1.28 text: "Verify the certificate and key used to encrypt API server traffic" audit: "grep -A7 servingInfo /etc/origin/master/master-config.yaml" tests: bin_op: and test_items: - flag: "certFile" compare: op: has value: "master.server.crt" set: true - flag: "keyFile" compare: op: has value: "master.server.key" set: true remediation: | Edit the Openshift master config file /etc/origin/master/master-config.yaml and set keyFile and certFile under servingInfo. servingInfo:  bindAddress: 0.0.0.0:8443   bindNetwork: tcp4 certFile: master.server.crt clientCA: ca.crt keyFile: master.server.key maxRequestsInFlight: 500 requestTimeoutSeconds: 3600 scored: true - id: 1.29 text: "Verify that the --client-ca-file argument is not set" audit: "grep client-ca-file /etc/origin/master/master-config.yaml" tests: test_items: - flag: "clientCA: ca.crt" set: false remediation: | Edit the Openshift master config file /etc/origin/master/master-config.yaml and set clientCA under servingInfo. servingInfo:  bindAddress: 0.0.0.0:8443   bindNetwork: tcp4 certFile: master.server.crt clientCA: ca.crt keyFile: master.server.key maxRequestsInFlight: 500 requestTimeoutSeconds: 3600 scored: true - id: 1.30 text: "Verify the CA used for communication with etcd" audit: "grep -A3 etcdClientInfo /etc/origin/master/master-config.yaml" tests: test_items: - flag: "ca" compare: op: has value: "master.etcd-ca.crt" set: true remediation: | Edit the Openshift master config file /etc/origin/master/master-config.yaml and set ca under etcdClientInfo. etcdClientInfo:   ca: master.etcd-ca.crt certFile: master.etcd-client.crt keyFile: master.etcd-client.key scored: true - id: 1.31 text: "Verify that the authorization-mode argument is not set" type: "skip" scored: true - id: 1.32 text: "Verify that the NodeRestriction admission controller is enabled" audit: "grep -A4 NodeRestriction /etc/origin/master/master-config.yaml" tests: bin_op: or test_items: - flag: "NodeRestriction" set: false - flag: "disable" compare: op: has value: "false" set: true remediation: | Edit the Openshift master config file /etc/origin/master/master-config.yaml and enable NodeRestriction ca under etcdClientInfo. NodeRestriction: configuration: kind: DefaultAdmissionConfig apiVersion: v1 disable: false scored: true - id: 1.33 text: "Configure encryption of data at rest in etcd datastore" audit: "grep -A1 experimental-encryption-provider-config /etc/origin/master/master-config.yaml" tests: test_items: - flag: "experimental-encryption-provider-config:" compare: op: has value: "experimental-encryption-provider-config:" set: true remediation: | Follow the instructions in the documentation to configure encryption. https://docs.openshift.com/container-platform/3.10/admin_guide/encrypting_data.html scored: true - id: 1.34 text: "Set the encryption provider to aescbc for etcd data at rest" audit: "grep -A1 experimental-encryption-provider-config /etc/origin/master/master-config.yaml | sed -n '2p' | awk '{ print $2 }' | xargs grep -A1 providers" tests: test_items: - flag: "aescbc:" compare: op: has value: "aescbc:" set: true remediation: | Edit the Openshift master config file /etc/origin/master/master-config.yaml and set aescbc as the first provider in encryption provider config. See https://docs.openshift.com/container-platform/3.10/admin_guide/encrypting_data.html. scored: true - id: 1.35 text: "Enable the EventRateLimit plugin" audit: "grep -A4 EventRateLimit /etc/origin/master/master-config.yaml" tests: test_items: - flag: "disable" compare: op: has value: "false" set: true remediation: | Follow the documentation to enable the EventRateLimit plugin. https://docs.openshift.com/container-platform/3.10/architecture/additional_concepts/admission_controllers.html#admission-controllers-general-admission-rules scored: true - id: 1.36 text: "Configure advanced auditing" audit: "grep AdvancedAuditing /etc/origin/master/master-config.yaml" tests: bin_op: or test_items: - flag: "AdvancedAuditing" compare: op: eq value: "true" set: true - flag: "AdvancedAuditing" set: false remediation: | Edit the Openshift master config file /etc/origin/master/master-config.yaml and enable AdvancedAuditing, kubernetesMasterConfig:  apiServerArguments: feature-gates: - AdvancedAuditing=true scored: true # Review 1.1.37 in Aquasec shared doc, the tests are net zero. - id: 1.37 text: "Adjust the request timeout argument for your cluster resources" audit: "grep request-timeout /etc/origin/master/master-config.yaml" type: "manual" remediation: | [Manual test] change the request-timeout value in the  /etc/origin/master/master-config.yaml scored: true - id: 2 text: "Scheduler" checks: - id: 2.1 text: "Verify that Scheduler profiling is not exposed to the web" type: "skip" scored: true - id: 3 text: "Controller Manager" checks: - id: 3.1 text: "Adjust the terminated-pod-gc-threshold argument as needed" audit: "grep terminated-pod-gc-threshold -A1 /etc/origin/master/master-config.yaml" tests: test_items: - flag: "terminated-pod-gc-threshold:" compare: op: has value: "12500" set: true remediation: | Edit the Openshift master config file /etc/origin/master/master-config.yaml and enable terminated-pod-gc-threshold. kubernetesMasterConfig:  controllerArguments:     terminated-pod-gc-threshold:    - true Enabling the "terminated-pod-gc-threshold" settings is optional. scored: true - id: 3.2 text: "Verify that Controller profiling is not exposed to the web" type: "skip" scored: true - id: 3.3 text: "Verify that the --use-service-account-credentials argument is set to true" audit: "grep -A2 use-service-account-credentials /etc/origin/master/master-config.yaml" tests: bin_op: or test_items: - flag: "use-service-account-credentials" set: false - flag: "true" compare: op: has value: "true" set: true remediation: | Edit the Openshift master config file /etc/origin/master/master-config.yaml and set use-service-account-credentials to true under controllerArguments section. kubernetesMasterConfig:  controllerArguments:     use-service-account-credentials:     - true scored: true # Review 3.4 - id: 3.4 text: "Verify that the --service-account-private-key-file argument is set as appropriate" audit: | grep -A9 serviceAccountConfig /etc/origin/master/master-config.yaml | grep privateKeyFile; grep -A2 service-account-private-key-file /etc/origin/master/master-config.yaml tests: bin_op: and test_items: - flag: "privateKeyFile: serviceaccounts.private.key" compare: op: has value: "privateKeyFile" - flag: "service-account-private-key-file" set: false remediation: Edit the Openshift master config file /etc/origin/master/master-config.yaml and remove service-account-private-key-file scored: true # Review 3.5 - id: 3.5 text: "Verify that the --root-ca-file argument is set as appropriate" audit: "/bin/sh -c 'grep root-ca-file /etc/origin/master/master-config.yaml; grep -A9 serviceAccountConfig /etc/origin/master/master-config.yaml'" tests: bin_op: and test_items: - flag: "root-ca-file=/etc/origin/master/ca-bundle.crt" compare: op: has value: "/etc/origin/master/ca-bundle.crt" set: true - flag: "masterCA" compare: op: has value: "ca-bundle.crt" set: true remediation: Reset to OpenShift defaults OpenShift starts kube-controller-manager with root-ca-file=/etc/origin/master/ca-bundle.crt by default.  OpenShift Advanced Installation creates this certificate authority and configuration without any configuration required. https://docs.openshift.com/container-platform/3.10/admin_guide/service_accounts.html" scored: true - id: 3.6 text: "Verify that Security Context Constraints are applied to Your Pods and Containers" type: "skip" scored: false - id: 3.7 text: "Manage certificate rotation" audit: "grep -B3 RotateKubeletServerCertificate=true /etc/origin/master/master-config.yaml" tests: test_items: - flag: "RotateKubeletServerCertificate" compare: op: eq value: "true" set: true remediation: If you decide not to enable the RotateKubeletServerCertificate feature, be sure to use the Ansible playbooks provided with the OpenShift installer to automate re-deploying certificates. scored: true - id: 4 text: "Configuration Files" checks: - id: 4.1 text: "Verify the OpenShift default permissions for the API server pod specification file" audit: "stat -c permissions=%a /etc/origin/node/pods/apiserver.yaml" tests: test_items: - flag: "permissions" compare: op: bitmask value: "600" set: true remediation: | Run the below command. chmod 600 /etc/origin/node/pods/apiserver.yaml scored: true - id: 4.2 text: "Verify the OpenShift default file ownership for the API server pod specification file" audit: "stat -c %U:%G /etc/origin/node/pods/apiserver.yaml" tests: test_items: - flag: "root:root" compare: op: eq value: "root:root" set: true remediation: | Run the below command on the master node. chown root:root /etc/origin/node/pods/apiserver.yaml scored: true - id: 4.3 text: "Verify the OpenShift default file permissions for the controller manager pod specification file" audit: "stat -c permissions=%a /etc/origin/node/pods/controller.yaml" tests: test_items: - flag: "permissions" compare: op: bitmask value: "600" set: true remediation: | Run the below command on the master node. chmod 600 /etc/origin/node/pods/controller.yaml scored: true - id: 4.4 text: "Verify the OpenShift default ownership for the controller manager pod specification file" audit: "stat -c %U:%G /etc/origin/node/pods/controller.yaml" tests: test_items: - flag: "root:root" compare: op: eq value: "root:root" set: true remediation: | Run the below command on the master node. chown root:root /etc/origin/node/pods/controller.yaml scored: true - id: 4.5 text: "Verify the OpenShift default permissions for the scheduler pod specification file" audit: "stat -c permissions=%a /etc/origin/node/pods/controller.yaml" tests: test_items: - flag: "permissions" compare: op: bitmask value: "600" set: true remediation: | Run the below command. chmod 600 stat -c permissions=%a /etc/origin/node/pods/controller.yaml scored: true - id: 4.6 text: "Verify the scheduler pod specification file ownership set by OpenShift" audit: "stat -c %u:%g /etc/origin/node/pods/controller.yaml" tests: test_items: - flag: "root:root" compare: op: eq value: "root:root" set: true remediation: | Run the below command on the master node. chown root:root /etc/origin/node/pods/controller.yaml scored: true - id: 4.7 text: "Verify the OpenShift default etcd pod specification file permissions" audit: "stat -c permissions=%a /etc/origin/node/pods/etcd.yaml" tests: test_items: - flag: "permissions" compare: op: bitmask value: "600" set: true remediation: | Run the below command. chmod 600 /etc/origin/node/pods/etcd.yaml scored: true - id: 4.8 text: "Verify the OpenShift default etcd pod specification file ownership" audit: "stat -c %U:%G /etc/origin/node/pods/etcd.yaml" tests: test_items: - flag: "root:root" compare: op: eq value: "root:root" set: true remediation: | Run the below command on the master node. chown root:root /etc/origin/node/pods/etcd.yaml scored: true - id: 4.9 text: "Verify the default OpenShift Container Network Interface file permissions" audit: "stat -c permissions=%a /etc/origin/openvswitch/ /etc/cni/net.d/" tests: test_items: - flag: "permissions" compare: op: bitmask value: "644" set: true remediation: | Run the below command. chmod 644 -R /etc/origin/openvswitch/ /etc/cni/net.d/ scored: true - id: 4.10 text: "Verify the default OpenShift Container Network Interface file ownership" audit: "stat -c %U:%G /etc/origin/openvswitch/ /etc/cni/net.d/" tests: test_items: - flag: "root:root" compare: op: eq value: "root:root" set: true remediation: | Run the below command on the master node. chown root:root /etc/origin/openvswitch/ /etc/cni/net.d/ scored: true - id: 4.11 text: "Verify the default OpenShift etcd data directory permissions" audit: "stat -c permissions=%a /var/lib/etcd" tests: test_items: - flag: "permissions" compare: op: bitmask value: "700" set: true remediation: | On the etcd server node, get the etcd data directory, passed as an argument --data-dir , from the below command: ps -ef | grep etcd Run the below command (based on the etcd data directory found above). For example, chmod 700 /var/lib/etcd scored: true - id: 4.12 text: "Verify the default OpenShift etcd data directory ownership" audit: "stat -c %U:%G /var/lib/etcd" tests: test_items: - flag: "etcd:etcd" compare: op: eq value: "etcd:etcd" set: true remediation: | Run the below command on the master node. chown etcd:etcd /var/lib/etcd scored: true - id: 4.13 text: "Verify the default OpenShift admin.conf file permissions" audit: "stat -c permissions=%a /etc/origin/master/admin.kubeconfig" tests: test_items: - flag: "permissions" compare: op: bitmask value: "644" set: true remediation: | Run the below command. chmod 644 /etc/origin/master/admin.kubeconfig" scored: true - id: 4.14 text: "Verify the default OpenShift admin.conf file ownership" audit: "stat -c %U:%G /etc/origin/master/admin.kubeconfig" tests: test_items: - flag: "root:root" compare: op: eq value: "root:root" set: true remediation: | Run the below command on the master node. chown root:root /etc/origin/master/admin.kubeconfig scored: true - id: 4.15 text: "Verify the default OpenShift scheduler.conf file permissions" audit: "stat -c permissions=%a /etc/origin/master/openshift-master.kubeconfig" tests: test_items: - flag: "permissions" compare: op: bitmask value: "644" set: true remediation: | Run the below command. chmod 644 /etc/origin/master/openshift-master.kubeconfig scored: true - id: 4.16 text: "Verify the default OpenShift scheduler.conf file ownership" audit: "stat -c %U:%G /etc/origin/master/openshift-master.kubeconfig" tests: test_items: - flag: "root:root" compare: op: eq value: "root:root" set: true remediation: | Run the below command on the master node. chown root:root /etc/origin/master/openshift-master.kubeconfig scored: true - id: 4.17 text: "Verify the default Openshift controller-manager.conf file permissions" audit: "stat -c permissions=%a /etc/origin/master/openshift-master.kubeconfig" tests: test_items: - flag: "permissions" compare: op: bitmask value: "644" set: true remediation: | Run the below command. chmod 644 /etc/origin/master/openshift-master.kubeconfig scored: true - id: 4.18 text: "Ensure that the controller-manager.conf file ownership is set to root:root (Scored)" audit: "stat -c %U:%G /etc/origin/master/openshift-master.kubeconfig" tests: test_items: - flag: "root:root" compare: op: eq value: "root:root" set: true remediation: | Run the below command on the master node. chown root:root /etc/origin/master/openshift-master.kubeconfig scored: true - id: 5 text: "Etcd" checks: - id: 5.1 text: "Verify the default OpenShift cert-file and key-file configuration" audit: "/bin/sh -c '/usr/local/bin/master-exec etcd etcd grep ETCD_CERT_FILE=/etc/etcd/server.crt /proc/1/environ; /usr/local/bin/master-exec etcd etcd grep etcd_key_file=/etc/etcd/server.key /proc/1/environ; grep ETCD_CERT_FILE=/etc/etcd/server.crt /etc/etcd/etcd.conf; grep ETCD_KEY_FILE=/etc/etcd/server.key /etc/etcd/etcd.conf'" tests: bin_op: and test_items: - flag: "Binary file /proc/1/environ matches" compare: op: has value: "Binary file /proc/1/environ matches" set: true - flag: "ETCD_CERT_FILE=/etc/etcd/server.crt" compare: op: has value: "ETCD_CERT_FILE=/etc/etcd/server.crt" set: true - flag: "ETCD_KEY_FILE=/etc/etcd/server.key" compare: op: has value: "ETCD_KEY_FILE=/etc/etcd/server.key" set: true remediation: | Reset to the OpenShift default configuration. scored: true - id: 5.2 text: "Verify the default OpenShift setting for the client-cert-auth argument" audit: "/bin/sh -c'/usr/local/bin/master-exec etcd etcd grep ETCD_CLIENT_CERT_AUTH=true /proc/1/environ; grep ETCD_CLIENT_CERT_AUTH /etc/etcd/etcd.conf'" tests: bin_op: and test_items: - flag: "Binary file /proc/1/environ matches" compare: op: has value: "Binary file /proc/1/environ matches" set: true - flag: "ETCD_CLIENT_CERT_AUTH=true" compare: op: has value: "ETCD_CLIENT_CERT_AUTH=true" set: true remediation: | Reset to the OpenShift default configuration. scored: true - id: 5.3 text: "Verify the OpenShift default values for etcd_auto_tls" audit: "/bin/sh -c '/usr/local/bin/master-exec etcd etcd grep ETCD_AUTO_TLS /proc/1/environ; grep ETCD_AUTO_TLS /etc/etcd/etcd.conf'" tests: bin_op: or test_items: - flag: "ETCD_AUTO_TLS=false" compare: op: has value: "ETCD_AUTO_TLS=false" set: true - flag: "#ETCD_AUTO_TLS" compare: op: has value: "#ETCD_AUTO_TLS" set: true remediation: | Reset to the OpenShift default configuration. scored: true - id: 5.4 text: "Verify the OpenShift default peer-cert-file and peer-key-file arguments for etcd" audit: "/bin/sh -c'/usr/local/bin/master-exec etcd etcd grep ETCD_PEER_CERT_FILE=/etc/etcd/peer.crt /proc/1/environ; /usr/local/bin/master-exec etcd etcd grep ETCD_PEER_KEY_FILE=/etc/etcd/peer.key /proc/1/environ; grep ETCD_PEER_CERT_FILE /etc/etcd/etcd.conf; grep ETCD_PEER_KEY_FILE /etc/etcd/etcd.conf'" tests: bin_op: and test_items: - flag: "Binary file /proc/1/environ matches" compare: op: has value: "Binary file /proc/1/environ matches" set: true - flag: "ETCD_PEER_CERT_FILE=/etc/etcd/peer.crt" compare: op: has value: "ETCD_PEER_CERT_FILE=/etc/etcd/peer.crt" set: true - flag: "ETCD_PEER_KEY_FILE=/etc/etcd/peer.key" compare: op: has value: "ETCD_PEER_KEY_FILE=/etc/etcd/peer.key" set: true remediation: | Reset to the OpenShift default configuration. scored: true - id: 5.5 text: "Verify the OpenShift default configuration for the peer-client-cert-auth" audit: "/bin/sh -c '/usr/local/bin/master-exec etcd etcd grep ETCD_PEER_CLIENT_CERT_AUTH=true /proc/1/environ; grep ETCD_PEER_CLIENT_CERT_AUTH /etc/etcd/etcd.conf'" tests: bin_op: and test_items: - flag: "Binary file /proc/1/environ matches" compare: op: has value: "Binary file /proc/1/environ matches" set: true - flag: "ETCD_PEER_CLIENT_CERT_AUTH=true" compare: op: has value: "ETCD_PEER_CLIENT_CERT_AUTH=true" set: true remediation: | Reset to the OpenShift default configuration. scored: true - id: 5.6 text: "Verify the OpenShift default configuration for the peer-auto-tls argument" audit: "/bin/sh -c '/usr/local/bin/master-exec etcd etcd grep ETCD_PEER_AUTO_TLS /proc/1/environ; grep ETCD_PEER_AUTO_TLS /etc/etcd/etcd.conf'" tests: bin_op: and test_items: - flag: "Binary file /proc/1/environ matches" compare: op: has value: "Binary file /proc/1/environ matches" set: true - flag: "#ETCD_PEER_AUTO_TLS=false" compare: op: has value: "#ETCD_PEER_AUTO_TLS=false" set: true remediation: | Reset to the OpenShift default configuration. scored: true - id: 5.7 text: "Optionally modify the wal-dir argument" type: "skip" scored: true - id: 5.8 text: "Optionally modify the max-wals argument" type: "skip" scored: true - id: 5.9 text: "Verify the OpenShift default configuration for the etcd Certificate Authority" audit: "openssl x509 -in /etc/origin/master/master.etcd-ca.crt -subject -issuer -noout | sed 's/@/ /'" tests: test_items: - flag: "issuer= /CN=etcd-signer" compare: op: has value: "issuer= /CN=etcd-signer" set: true remediation: | Reset to the OpenShift default configuration. scored: false - id: 6 text: "General Security Primitives" checks: - id: 6.1 text: "Ensure that the cluster-admin role is only used where required" type: "manual" remediation: | [Manual test] Review users, groups, serviceaccounts bound to cluster-admin: oc get clusterrolebindings | grep cluster-admin Review users and groups bound to cluster-admin and decide whether they require such access. Consider creating least-privilege roles for users and service accounts scored: false - id: 6.2 text: "Verify Security Context Constraints as in use" type: "manual" remediation: | [Manual test] Review Security Context Constraints: oc get scc Use OpenShift's Security Context Constraint feature, which has been contributed to Kubernetes as Pod Security Policies. PSPs are still beta in Kubernetes 1.10. OpenShift ships with two SCCs: restricted and privileged. The two default SCCs will be created when the master is started. The restricted SCC is granted to all authenticated users by default. https://docs.openshift.com/container-platform/3.10/admin_guide/manage_scc.html" scored: false - id: 6.3 text: "Use OpenShift projects to maintain boundaries between resources" type: "manual" remediation: | [Manual test] Review projects: oc get projects scored: false - id: 6.4 text: "Create network segmentation using the Multi-tenant plugin or Network Policies" type: "manual" remediation: | [Manual test] Verify on masters the plugin being used: grep networkPluginName /etc/origin/master/master-config.yaml OpenShift provides multi-tenant networking isolation (using Open vSwich and vXLAN), to segregate network traffic between containers belonging to different tenants (users or applications) while running on a shared cluster. Red Hat also works with 3rd-party SDN vendors to provide the same level of capabilities integrated with OpenShift. OpenShift SDN is included a part of OpenShift subscription. OpenShift supports Kubernetes NetworkPolicy. Administrator must configure NetworkPolicies if desired. https://docs.openshift.com/container-platform/3.10/architecture/networking/sdn.html#architecture-additional-concepts-sdn Ansible Inventory variable: os_sdn_network_plugin_name: https://docs.openshift.com/container-platform/3.10/install/configuring_inventory_file.html scored: false - id: 6.5 text: "Enable seccomp and configure custom Security Context Constraints" type: "manual" remediation: | [Manual test] Verify SCCs that have been configured with seccomp: oc get scc -ocustom-columns=NAME:.metadata.name,SECCOMP-PROFILES:.seccompProfiles OpenShift does not enable seccomp by default. To configure seccomp profiles that are applied to pods run by the SCC, follow the instructions in the documentation: https://docs.openshift.com/container-platform/3.9/admin_guide/seccomp.html#admin-guide-seccomp scored: false - id: 6.6 text: "Review Security Context Constraints" type: "manual" remediation: | [Manual test] Review SCCs: oc describe scc Use OpenShift's Security Context Constraint feature, which has been contributed to Kubernetes as Pod Security Policies. PSPs are still beta in Kubernetes 1.10. OpenShift ships with two SCCs: restricted and privileged. The two default SCCs will be created when the master is started. The restricted SCC is granted to all authenticated users by default. All pods are run under the restricted SCC by default. Running a pod under any other SCC requires an account with cluster admin capabilities to grant access for the service account. SecurityContextConstraints limit what securityContext is applied to pods and containers. https://docs.openshift.com/container-platform/3.10/admin_guide/manage_scc.html scored: false - id: 6.7 text: "Manage Image Provenance using ImagePolicyWebhook admission controller" type: "manual" remediation: | [Manual test] Review imagePolicyConfig in /etc/origin/master/master-config.yaml. scored: false - id: 6.8 text: "Configure Network policies as appropriate" type: "manual" remediation: | [Manual test] If ovs-networkplugin is used, review network policies: oc get networkpolicies OpenShift supports Kubernetes NetworkPolicy via ovs-networkpolicy plugin. If choosing ovs-multitenant plugin, each namespace is isolated in its own netnamespace by default. scored: false - id: 6.9 text: "Use Security Context Constraints as compensating controls for privileged containers" type: "manual" remediation: | [Manual test] 1) Determine all sccs allowing privileged containers: oc get scc -ocustom-columns=NAME:.metadata.name,ALLOWS_PRIVILEGED:.allowPrivilegedContainer 2) Review users and groups assigned to sccs allowing priviliged containers: oc describe sccs Use OpenShift's Security Context Constraint feature, which has been contributed to Kubernetes as Pod Security Policies. PSPs are still beta in Kubernetes 1.10. OpenShift ships with two SCCs: restricted and privileged. The two default SCCs will be created when the master is started. The restricted SCC is granted to all authenticated users by default. Similar scenarios are documented in the SCC documentation, which outlines granting SCC access to specific serviceaccounts. Administrators may create least-restrictive SCCs based on individual container needs. For example, if a container only requires running as the root user, the anyuid SCC can be used, which will not expose additional access granted by running privileged containers. https://docs.openshift.com/container-platform/3.10/admin_guide/manage_scc.html scored: false