The Docker image may now exit to shell

This patch updates the Dockerfile to use an ENTRYPOINT command that points to the file "/cncf/provision.sh" instead of the CMD directive. The provision script has been updated to shell exec "/bin/bash" if the first argument to the script is "shell". Otherwise the provision script reads all of the same environment variables that used to be passed to the script as command-line arguments from the Dockerfile. Additionally, the script still accepts the same command-line arguments in the same order. Since environment variables cannot be used in a Dockerfile's CMD or ENTRYPOINT directives without using them in a second "sh" (or other shell) directive, they are no longer given to ENTRYPOINT or CMD as the default values for the provision script. Instead, as documented above, the values are parsed directly in the provision script.
akutz · Jun 15, 2018 · 85f69f2 · 85f69f2
1 parent b49a6b9
commit 85f69f2
Show file tree

Hide file tree

Showing 2 changed files with 67 additions and 50 deletions.
diff --git a/Dockerfile b/Dockerfile
@@ -110,4 +110,4 @@ COPY worker_templates-v1.10.0/ /cncf/worker_templates-v1.10.0/
 RUN chmod +x /cncf/provision.sh
 WORKDIR /cncf/
 
-CMD ["bash", "-c", "/cncf/provision.sh ${CLOUD}-${COMMAND} ${NAME} ${BACKEND} ${DATA_FOLDER}"]
+ENTRYPOINT ["/cncf/provision.sh"]
diff --git a/provision.sh b/provision.sh
@@ -1,8 +1,22 @@
 #!/bin/bash
 # Usage
 #
+# provision.sh shell
 # provision.sh <provider>-<command> <name> <config-backend>
 #
+
+# If the first argument is "shell" then exec into a shell and abandon
+# the rest of the script.
+if [ "$1" = "shell" ]; then exec /bin/bash; fi
+
+# Initialize the configuration properties based on the command-line
+# arguments OR on the corresponding environment variables.
+CLOUD_CMD=$CLOUD-$COMMAND
+if [ ! "$1" = "" ]; then CLOUD_CMD=$1; fi
+if [ ! "$2" = "" ]; then NAME=$2; fi
+if [ ! "$3" = "" ]; then BACKEND=$3; fi
+if [ ! "$4" = "" ]; then DATA_FOLDER=$4; fi
+
 DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )"
 _retry() {
     [ -z "${2}" ] && return 1
@@ -15,8 +29,8 @@ RED='\033[0;31m'
 NC='\033[0m' # No Color
 
 # Setup Enviroment Using $NAME
-export TF_VAR_name="$2"
-export TF_VAR_data_dir=$(pwd)/data/"$4"
+export TF_VAR_name="$NAME"
+export TF_VAR_data_dir=$(pwd)/data/"$DATA_FOLDER"
 export TF_VAR_packet_api_key=${PACKET_AUTH_TOKEN}
 
 # Configure Artifacts
@@ -30,7 +44,7 @@ fi
 
 
 if [ ! -e $ETCD_ARTIFACT ] ; then
-  export TF_VAR_etcd_artifact=$ETCD_ARTIFACT
+    export TF_VAR_etcd_artifact=$ETCD_ARTIFACT
 fi
 
 if [ ! -e $KUBE_APISERVER_ARTIFACT ] ; then
@@ -59,9 +73,9 @@ fi
 mkdir -p $TF_VAR_data_dir
 
 # Run CMD
-if [ "$1" = "aws-deploy" ] ; then
+if [ "$CLOUD_CMD" = "aws-deploy" ] ; then
     cd ${DIR}/aws
-    if [ "$3" = "s3" ]; then
+    if [ "$BACKEND" = "s3" ]; then
         cp ../s3-backend.tf .
     terraform init \
               -backend-config "bucket=${AWS_BUCKET}" \
@@ -70,7 +84,7 @@ if [ "$1" = "aws-deploy" ] ; then
     # ensure kubeconfig is written to disk on infrastructure refresh
     terraform taint -module=kubeconfig null_resource.kubeconfig || true
     time terraform apply -auto-approve ${DIR}/aws
-    elif [ "$3" = "file" ]; then
+    elif [ "$BACKEND" = "file" ]; then
         cp ../file-backend.tf .
         terraform init \
                   -backend-config "path=/cncf/data/${TF_VAR_name}/terraform.tfstate"
@@ -85,29 +99,29 @@ if [ "$1" = "aws-deploy" ] ; then
     _retry "❤ Ensure that ClusterRoles are available" kubectl get ClusterRole.v1.rbac.authorization.k8s.io
     _retry "❤ Ensure that ClusterRoleBindings are available" kubectl get ClusterRoleBinding.v1.rbac.authorization.k8s.io
 
-elif [ "$1" = "aws-destroy" ] ; then
+elif [ "$CLOUD_CMD" = "aws-destroy" ] ; then
       cd ${DIR}/aws
-      if [ "$3" = "s3" ]; then
+      if [ "$BACKEND" = "s3" ]; then
           cp ../s3-backend.tf .
           terraform init \
                     -backend-config "bucket=${AWS_BUCKET}" \
                     -backend-config "key=aws-${TF_VAR_name}" \
                     -backend-config "region=${AWS_DEFAULT_REGION}"
     time terraform destroy -force ${DIR}/aws
-      elif [ "$3" = "file" ]; then
+      elif [ "$BACKEND" = "file" ]; then
           cp ../file-backend.tf .
           terraform init \
                     -backend-config "path=/cncf/data/${TF_VAR_name}/terraform.tfstate"
           time terraform destroy -force ${DIR}/aws
 
        fi
 
-elif [ "$1" = "azure-deploy" ] ; then
+elif [ "$CLOUD_CMD" = "azure-deploy" ] ; then
     # There are some dependency issues around cert,sshkey,k8s_cloud_config, and dns
     # since they use files on disk that are created on the fly
     # should probably move these to data resources
     cd ${DIR}/azure
-    if [ "$3" = "s3" ]; then
+    if [ "$BACKEND" = "s3" ]; then
         cp ../s3-backend.tf .
     terraform init \
               -backend-config "bucket=${AWS_BUCKET}" \
@@ -120,7 +134,7 @@ elif [ "$1" = "azure-deploy" ] ; then
         terraform apply -target module.network.azurerm_subnet.cncf -auto-approve ${DIR}/azure || true && \
         time terraform apply -auto-approve ${DIR}/azure
 
-    elif [ "$3" = "file" ]; then
+    elif [ "$BACKEND" = "file" ]; then
         cp ../file-backend.tf .
         terraform init \
                   -backend-config "path=/cncf/data/${TF_VAR_name}/terraform.tfstate"
@@ -139,44 +153,45 @@ elif [ "$1" = "azure-deploy" ] ; then
     _retry "❤ Ensure that ClusterRoleBindings are available" kubectl get ClusterRoleBinding.v1.rbac.authorization.k8s.io
 
 
-elif [ "$1" = "azure-destroy" ] ; then
+elif [ "$CLOUD_CMD" = "azure-destroy" ] ; then
     cd ${DIR}/azure
-    if [ "$3" = "s3" ]; then
+    if [ "$BACKEND" = "s3" ]; then
         cp ../s3-backend.tf .
     terraform init \
               -backend-config "bucket=${AWS_BUCKET}" \
               -backend-config "key=azure-${TF_VAR_name}" \
               -backend-config "region=${AWS_DEFAULT_REGION}"
     time terraform destroy -force ${DIR}/azure || true
 
-    elif [ "$3" = "file" ]; then
+    elif [ "$BACKEND" = "file" ]; then
         cp ../file-backend.tf .
         terraform init \
                   -backend-config "path=/cncf/data/${TF_VAR_name}/terraform.tfstate"
     time terraform destroy -force ${DIR}/azure || true
     fi
 
 # Begin OpenStack
-elif [[ "$1" = "openstack-deploy" || "$1" = "openstack-destroy" ]] ; then
+elif [[ "$CLOUD_CMD" = "openstack-deploy" || \
+        "$CLOUD_CMD" = "openstack-destroy" ]] ; then
     cd ${DIR}/openstack
 
     # initialize based on the config type
-    if [ "$3" = "s3" ] ; then
+    if [ "$BACKEND" = "s3" ] ; then
         cp ../s3-backend.tf .
         terraform init \
             -backend-config "bucket=${AWS_BUCKET}" \
             -backend-config "key=openstack-${TF_VAR_name}" \
             -backend-config "region=${AWS_DEFAULT_REGION}"
-    elif [ "$3" = "file" ] ; then
+    elif [ "$BACKEND" = "file" ] ; then
         cp ../file-backend.tf .
         terraform init -backend-config "path=/cncf/data/${TF_VAR_NAME}/terraform.tfstate"
     fi
 
     # deploy/destroy implementations
-    if [ "$1" = "openstack-deploy" ] ; then
+    if [ "$CLOUD_CMD" = "openstack-deploy" ] ; then
         terraform taint -module=kubeconfig null_resource.kubeconfig || true
         time terraform apply -auto-approve ${DIR}/openstack
-    elif [ "$1" = "openstack-destroy" ] ; then
+    elif [ "$CLOUD_CMD" = "openstack-destroy" ] ; then
         time terraform destroy -force ${DIR}/openstack || true
         # Exit after destroying resources as further commands cause hang
         exit
@@ -190,9 +205,9 @@ elif [[ "$1" = "openstack-deploy" || "$1" = "openstack-destroy" ]] ; then
 
 # End OpenStack
 
-elif [ "$1" = "packet-deploy" ] ; then
+elif [ "$CLOUD_CMD" = "packet-deploy" ] ; then
     cd ${DIR}/packet
-    if [ "$3" = "s3" ]; then
+    if [ "$BACKEND" = "s3" ]; then
         cp ../s3-backend.tf .
     terraform init \
               -backend-config "bucket=${AWS_BUCKET}" \
@@ -202,7 +217,7 @@ elif [ "$1" = "packet-deploy" ] ; then
     terraform taint -module=kubeconfig null_resource.kubeconfig || true ${DIR}/packet
     time terraform apply -auto-approve ${DIR}/packet
 
-    elif [ "$3" = "file" ]; then
+    elif [ "$BACKEND" = "file" ]; then
         cp ../file-backend.tf .
         terraform init \
                   -backend-config "path=/cncf/data/${TF_VAR_name}/terraform.tfstate" 
@@ -217,26 +232,26 @@ fi
     _retry "❤ Ensure that ClusterRoles are available" kubectl get ClusterRole.v1.rbac.authorization.k8s.io
     _retry "❤ Ensure that ClusterRoleBindings are available" kubectl get ClusterRoleBinding.v1.rbac.authorization.k8s.io
 
-elif [ "$1" = "packet-destroy" ] ; then
+elif [ "$CLOUD_CMD" = "packet-destroy" ] ; then
      cd ${DIR}/packet
-     if [ "$3" = "s3" ]; then
+     if [ "$BACKEND" = "s3" ]; then
          cp ../s3-backend.tf .
     terraform init \
               -backend-config "bucket=${AWS_BUCKET}" \
               -backend-config "key=packet-${TF_VAR_name}" \
               -backend-config "region=${AWS_DEFAULT_REGION}"
     time terraform destroy -force ${DIR}/packet
 
-elif [ "$3" = "file" ]; then
+elif [ "$BACKEND" = "file" ]; then
          cp ../file-backend.tf .
          terraform init \
                    -backend-config "path=/cncf/data/${TF_VAR_name}/terraform.tfstate"
          time terraform destroy -force ${DIR}/packet
 fi
 
-elif [ "$1" = "gce-deploy" ] ; then
+elif [ "$CLOUD_CMD" = "gce-deploy" ] ; then
     cd ${DIR}/gce
-    if [ "$3" = "s3" ]; then
+    if [ "$BACKEND" = "s3" ]; then
         cp ../s3-backend.tf .
     terraform init \
               -backend-config "bucket=${AWS_BUCKET}" \
@@ -247,7 +262,7 @@ elif [ "$1" = "gce-deploy" ] ; then
     time terraform apply -target module.vpc.google_compute_subnetwork.cncf -auto-approve ${DIR}/gce
     time terraform apply -auto-approve ${DIR}/gce
 
-elif [ "$3" = "file" ]; then
+elif [ "$BACKEND" = "file" ]; then
         cp ../file-backend.tf .
         terraform init \
                   -backend-config "path=/cncf/data/${TF_VAR_name}/terraform.tfstate"
@@ -263,9 +278,9 @@ elif [ "$3" = "file" ]; then
     _retry "❤ Ensure that ClusterRoles are available" kubectl get ClusterRole.v1.rbac.authorization.k8s.io
     _retry "❤ Ensure that ClusterRoleBindings are available" kubectl get ClusterRoleBinding.v1.rbac.authorization.k8s.io
 
-elif [ "$1" = "gce-destroy" ] ; then
+elif [ "$CLOUD_CMD" = "gce-destroy" ] ; then
     cd ${DIR}/gce
-    if [ "$3" = "s3" ]; then
+    if [ "$BACKEND" = "s3" ]; then
         cp ../s3-backend.tf .
     terraform init \
               -backend-config "bucket=${AWS_BUCKET}" \
@@ -274,7 +289,7 @@ elif [ "$1" = "gce-destroy" ] ; then
     time terraform destroy -force ${DIR}/gce || true # Allow to Fail and clean up network on next step
     time terraform destroy -force -target module.vpc.google_compute_subnetwork.cncf ${DIR}/gce
     time terraform destroy -force -target module.vpc.google_compute_network.cncf ${DIR}/gce
-elif [ "$3" = "file" ]; then
+elif [ "$BACKEND" = "file" ]; then
         cp ../file-backend.tf .
         terraform init \
                   -backend-config "path=/cncf/data/${TF_VAR_name}/terraform.tfstate"
@@ -283,17 +298,17 @@ elif [ "$3" = "file" ]; then
         time terraform destroy -force -target module.vpc.google_compute_network.cncf ${DIR}/gce
 fi
 
-elif [ "$1" = "gke-deploy" ] ; then
+elif [ "$CLOUD_CMD" = "gke-deploy" ] ; then
 cd ${DIR}/gke
-if [ "$3" = "s3" ]; then
+if [ "$BACKEND" = "s3" ]; then
     cp ../s3-backend.tf .
     terraform init \
               -backend-config "bucket=${AWS_BUCKET}" \
               -backend-config "key=gke-${TF_VAR_name}" \
               -backend-config "region=${AWS_DEFAULT_REGION}"
     time terraform apply -target module.vpc -auto-approve ${DIR}/gke && \
     time terraform apply -auto-approve ${DIR}/gke
-elif [ "$3" = "file" ]; then
+elif [ "$BACKEND" = "file" ]; then
     cp ../file-backend.tf .
     terraform init \
               -backend-config "path=/cncf/data/${TF_VAR_name}/terraform.tfstate"
@@ -310,9 +325,9 @@ fi
     _retry "❤ Trying to connect to cluster with kubectl" kubectl cluster-info 
     kubectl cluster-info
 
-elif [ "$1" = "gke-destroy" ] ; then
+elif [ "$CLOUD_CMD" = "gke-destroy" ] ; then
 cd ${DIR}/gke
-if [ "$3" = "s3" ]; then
+if [ "$BACKEND" = "s3" ]; then
     cp ../s3-backend.tf .
     terraform init \
               -backend-config "bucket=${AWS_BUCKET}" \
@@ -324,7 +339,7 @@ if [ "$3" = "s3" ]; then
     time terraform destroy -force -target module.vpc.google_compute_network.cncf ${DIR}/gke || true 
     time terraform destroy -force ${DIR}/gke || true
 
-elif [ "$3" = "file" ]; then
+elif [ "$BACKEND" = "file" ]; then
     cp ../file-backend.tf .
     terraform init \
               -backend-config "path=/cncf/data/${TF_VAR_name}/terraform.tfstate" 
@@ -335,9 +350,9 @@ time terraform destroy -force ${DIR}/gke || true
 fi
 
 
-elif [ "$1" = "ibmcloud-deploy" ] ; then
+elif [ "$CLOUD_CMD" = "ibmcloud-deploy" ] ; then
 cd ${DIR}/ibm
-if [ "$3" = "s3" ]; then
+if [ "$BACKEND" = "s3" ]; then
     cp ../s3-backend.tf .
     terraform init \
               -backend-config "bucket=${AWS_BUCKET}" \
@@ -346,7 +361,7 @@ if [ "$3" = "s3" ]; then
     # ensure kubeconfig is written to disk on infrastructure refresh
     terraform taint null_resource.kubeconfig || true
     time terraform apply -auto-approve ${DIR}/ibm
-elif [ "$3" = "file" ]; then
+elif [ "$BACKEND" = "file" ]; then
     cp ../file-backend.tf .
     terraform init \
               -backend-config "path=/cncf/data/${TF_VAR_name}/terraform.tfstate"
@@ -362,25 +377,27 @@ fi
     _retry "❤ Installing Helm" helm init
     kubectl rollout status -w deployment/tiller-deploy --namespace=kube-system
 
-elif [ "$1" = "ibmcloud-destroy" ] ; then
+elif [ "$CLOUD_CMD" = "ibmcloud-destroy" ] ; then
 cd ${DIR}/ibm
-if [ "$3" = "s3" ]; then
+if [ "$BACKEND" = "s3" ]; then
     cp ../s3-backend.tf .
     terraform init \
               -backend-config "bucket=${AWS_BUCKET}" \
               -backend-config "key=ibm-${TF_VAR_name}" \
               -backend-config "region=${AWS_DEFAULT_REGION}"
     time terraform destroy -force ${DIR}/ibm
 
-elif [ "$3" = "file" ]; then
+elif [ "$BACKEND" = "file" ]; then
     cp ../file-backend.tf .
     terraform init \
               -backend-config "path=/cncf/data/${TF_VAR_name}/terraform.tfstate" 
 time terraform destroy -force ${DIR}/ibm
 fi
 
 # Begin vSphere
-elif [[ "$1" = "vsphere-deploy" || "$1" = "vsphere-destroy" ]] ; then
+elif [[ "$CLOUD_CMD" = "vsphere-deploy" || \
+        "$CLOUD_CMD" = "vsphere-destroy" ]] ; then
+
     cd ${DIR}/vsphere
 
     export TF_VAR_vsphere_server=$VSPHERE_SERVER
@@ -391,22 +408,22 @@ elif [[ "$1" = "vsphere-deploy" || "$1" = "vsphere-destroy" ]] ; then
     export TF_VAR_vsphere_aws_region=$VSPHERE_AWS_REGION
 
     # initialize based on the config type
-    if [ "$3" = "s3" ] ; then
+    if [ "$BACKEND" = "s3" ] ; then
         cp ../s3-backend.tf .
         terraform init \
             -backend-config "bucket=${AWS_BUCKET}" \
             -backend-config "key=vsphere-${TF_VAR_name}" \
             -backend-config "region=${AWS_DEFAULT_REGION}"
-    elif [ "$3" = "file" ] ; then
+    elif [ "$BACKEND" = "file" ] ; then
         cp ../file-backend.tf .
         terraform init -backend-config "path=/cncf/data/${TF_VAR_name}/terraform.tfstate"
     fi
 
     # deploy/destroy implementations
-    if [ "$1" = "vsphere-deploy" ] ; then
+    if [ "$CLOUD_CMD" = "vsphere-deploy" ] ; then
         terraform taint -module=kubeconfig null_resource.kubeconfig || true
         time terraform apply -auto-approve ${DIR}/vsphere
-    elif [ "$1" = "vsphere-destroy" ] ; then
+    elif [ "$CLOUD_CMD" = "vsphere-destroy" ] ; then
         time terraform destroy -force ${DIR}/vsphere || true
         # Exit after destroying resources as further commands cause hang
         exit