feat: add IBM Cloud Example

ibm/terraform/main.tf

@@ -0,0 +1,23 @@
+module "resource-group" {
+  source  = "terraform-ibm-modules/resource-group/ibm"
+  version = "1.1.5"
+
+  resource_group_name = "qdrant-example-rg"
+}
+
+module "cluster" {
+  source  = "./modules/cluster"
+
+#  source  = "terraform-ibm-modules/base-ocp-vpc/ibm"
+#  version = "3.18.3"
+
+  ibmcloud_api_key     = var.ibmcloud_api_key
+  cluster_name         = var.cluster_name
+  resource_group_id    = module.resource-group.resource_group_id
+  region               = var.region
+  force_delete_storage = true
+  vpc_id               = module.vpc.vpc_id
+  vpc_subnets          = module.vpc.subnet_detail_map
+  worker_pools         = var.worker_pools
+}
+

ibm/terraform/modules/cluster/kconfig/.gitignore

@@ -0,0 +1,6 @@
+# Ignore everything
+*
+
+# But not these files...
+!.gitignore
+!README.md

ibm/terraform/modules/cluster/outputs.tf

@@ -0,0 +1,86 @@
+##############################################################################
+# Outputs
+##############################################################################
+
+output "cluster_id" {
+  description = "ID of cluster created"
+  value       = var.ignore_worker_pool_size_changes ? ibm_container_vpc_cluster.autoscaling_cluster[0].id : ibm_container_vpc_cluster.cluster[0].id
+  depends_on  = [null_resource.confirm_network_healthy]
+}
+
+output "cluster_name" {
+  description = "Name of the created cluster"
+  value       = var.ignore_worker_pool_size_changes ? ibm_container_vpc_cluster.autoscaling_cluster[0].name : ibm_container_vpc_cluster.cluster[0].name
+  depends_on  = [null_resource.confirm_network_healthy]
+}
+
+output "cluster_crn" {
+  description = "CRN for the created cluster"
+  value       = var.ignore_worker_pool_size_changes ? ibm_container_vpc_cluster.autoscaling_cluster[0].crn : ibm_container_vpc_cluster.cluster[0].crn
+  depends_on  = [null_resource.confirm_network_healthy]
+}
+
+output "workerpools" {
+  description = "Worker pools created"
+  value       = var.ignore_worker_pool_size_changes ? ibm_container_vpc_worker_pool.autoscaling_pool : ibm_container_vpc_worker_pool.pool
+}
+
+output "ocp_version" {
+  description = "Openshift Version of the cluster"
+  value       = var.ignore_worker_pool_size_changes ? ibm_container_vpc_cluster.autoscaling_cluster[0].kube_version : ibm_container_vpc_cluster.cluster[0].kube_version
+}
+
+output "cos_crn" {
+  description = "CRN of the COS instance"
+  value       = var.ignore_worker_pool_size_changes ? ibm_container_vpc_cluster.autoscaling_cluster[0].cos_instance_crn : ibm_container_vpc_cluster.cluster[0].cos_instance_crn
+}
+
+output "vpc_id" {
+  description = "ID of the clusters VPC"
+  value       = var.ignore_worker_pool_size_changes ? ibm_container_vpc_cluster.autoscaling_cluster[0].vpc_id : ibm_container_vpc_cluster.cluster[0].vpc_id
+}
+
+output "region" {
+  description = "Region cluster is deployed in"
+  value       = var.region
+}
+
+output "resource_group_id" {
+  description = "Resource group ID the cluster is deployed in"
+  value       = var.ignore_worker_pool_size_changes ? ibm_container_vpc_cluster.autoscaling_cluster[0].resource_group_id : ibm_container_vpc_cluster.cluster[0].resource_group_id
+}
+
+output "ingress_hostname" {
+  description = "Ingress hostname"
+  value       = var.ignore_worker_pool_size_changes ? ibm_container_vpc_cluster.autoscaling_cluster[0].ingress_hostname : ibm_container_vpc_cluster.cluster[0].ingress_hostname
+}
+
+output "private_service_endpoint_url" {
+  description = "Private service endpoint URL"
+  value       = var.ignore_worker_pool_size_changes ? ibm_container_vpc_cluster.autoscaling_cluster[0].private_service_endpoint_url : ibm_container_vpc_cluster.cluster[0].private_service_endpoint_url
+}
+
+output "public_service_endpoint_url" {
+  description = "Public service endpoint URL"
+  value       = var.ignore_worker_pool_size_changes ? ibm_container_vpc_cluster.autoscaling_cluster[0].public_service_endpoint_url : ibm_container_vpc_cluster.cluster[0].public_service_endpoint_url
+}
+
+output "master_url" {
+  description = "The URL of the Kubernetes master."
+  value       = var.ignore_worker_pool_size_changes ? ibm_container_vpc_cluster.autoscaling_cluster[0].master_url : ibm_container_vpc_cluster.cluster[0].master_url
+}
+
+output "kms_config" {
+  description = "KMS configuration details"
+  value       = var.kms_config
+}
+
+output "operating_system" {
+  description = "The operating system of the workers in the default worker pool."
+  value       = var.ignore_worker_pool_size_changes ? ibm_container_vpc_cluster.autoscaling_cluster[0].operating_system : ibm_container_vpc_cluster.cluster[0].operating_system
+}
+
+output "master_status" {
+  description = "The status of the Kubernetes master."
+  value       = var.ignore_worker_pool_size_changes ? ibm_container_vpc_cluster.autoscaling_cluster[0].master_status : ibm_container_vpc_cluster.cluster[0].master_status
+}

ibm/terraform/modules/cluster/scripts/confirm_network_healthy.sh

@@ -0,0 +1,96 @@
+#!/bin/bash
+
+set -e
+
+function run_checks() {
+
+  last_attempt=$1
+  namespace=calico-system
+
+  MAX_ATTEMPTS=10
+  attempt=0
+  PODS=()
+  while [ $attempt -lt $MAX_ATTEMPTS ]; do
+    # Get list of calico-node pods (There will be 1 pod per worker node)
+    if while IFS='' read -r line; do PODS+=("$line"); done < <(kubectl get pods -n "${namespace}" | grep calico-node | cut -f1 -d ' '); then
+      if [ ${#PODS[@]} -eq 0 ]; then
+        echo "No calico-node pods found. Retrying in 10s. (Attempt $((attempt+1)) / $MAX_ATTEMPTS)"
+        sleep 10
+        ((attempt=attempt+1))
+      else
+        # Pods found, break out of loop
+        break
+      fi
+    else
+      echo "Error getting calico-node pods. Retrying in 10s. (Attempt $((attempt+1)) / $MAX_ATTEMPTS)"
+      sleep 10
+      ((attempt=attempt+1))
+    fi
+  done
+
+  if [ ${#PODS[@]} -eq 0 ]; then
+    echo "No calico-node pods found after $MAX_ATTEMPTS attempts. Exiting."
+    exit 1
+  fi
+
+  # Iterate through pods to check health
+  healthy=true
+  for pod in "${PODS[@]}"; do
+    command="kubectl logs ${pod} -n ${namespace} --tail=0"
+    # If it is the last attempt then print the output
+    if [ "${last_attempt}" == true ]; then
+      node=$(kubectl get pod "$pod" -n "${namespace}" -o=jsonpath='{.spec.nodeName}')
+      echo "Checking node: $node"
+      if ! ${command}; then
+        healthy=false
+      else
+        echo "OK"
+      fi
+    # Otherwise redirect output to /dev/null
+    else
+      if ! ${command} &> /dev/null; then
+        healthy=false
+      fi
+    fi
+  done
+
+  if [ "$healthy" == "false" ]; then
+    return 1
+  else
+    return 0
+  fi
+
+}
+
+counter=0
+number_retries=40
+retry_wait_time=60
+
+echo "Running script to ensure kube master can communicate with all worker nodes.."
+
+while [ ${counter} -le ${number_retries} ]; do
+
+  # Determine if it is last attempt
+  last_attempt=false
+  if [ "${counter}" -eq ${number_retries} ]; then
+    last_attempt=true
+  fi
+
+  ((counter=counter+1))
+  if ! run_checks ${last_attempt}; then
+    if [ "${counter}" -gt ${number_retries} ]; then
+      echo "Maximum attempts reached, giving up."
+      echo
+      echo "Found kube master is unable to communicate with one or more of its workers."
+      echo "Please create a support issue with IBM Cloud and include the error message."
+      exit 1
+    else
+      echo "Retrying in ${retry_wait_time}s. (Retry attempt ${counter} / ${number_retries})"
+      sleep ${retry_wait_time}
+    fi
+  else
+    break
+  fi
+done
+
+echo "Success! Master can communicate with all worker nodes."

ibm/terraform/modules/cluster/scripts/get_config_map_status.sh

@@ -0,0 +1,23 @@
+#!/bin/bash
+
+set -e
+
+CONFIGMAP_NAME="iks-ca-configmap"
+NAMESPACE="kube-system"
+COUNTER=0
+MAX_ATTEMPTS=40
+
+while [[ $COUNTER -lt $MAX_ATTEMPTS ]] && ! kubectl get configmap $CONFIGMAP_NAME -n $NAMESPACE &>/dev/null; do
+  COUNTER=$((COUNTER + 1))
+  echo "Attempt $COUNTER: ConfigMap '$CONFIGMAP_NAME' not found in namespace '$NAMESPACE', retrying..."
+  sleep 60
+done
+
+if [[ $COUNTER -eq $MAX_ATTEMPTS ]]; then
+  echo "ConfigMap '$CONFIGMAP_NAME' did not become available within $MAX_ATTEMPTS attempts."
+  #  Output for debugging
+  kubectl get configmaps -n $NAMESPACE
+  exit 1
+else
+  echo "ConfigMap '$CONFIGMAP_NAME' is now available." >&2
+fi

ibm/terraform/modules/cluster/scripts/reset_iks_api_key.sh

@@ -0,0 +1,69 @@
+#!/bin/bash
+
+set -euo pipefail
+
+REGION="$1"
+RESOURCE_GROUP_ID="$2"
+APIKEY_KEY_NAME="containers-kubernetes-key"
+
+# Expects the environment variable $IBMCLOUD_API_KEY to be set
+if [[ -z "${IBMCLOUD_API_KEY}" ]]; then
+    echo "API key must be set with IBMCLOUD_API_KEY environment variable" >&2
+    exit 1
+fi
+
+if [[ -z "${REGION}" ]]; then
+    echo "Region must be passed as first input script argument" >&2
+    exit 1
+fi
+
+if [[ -z "${RESOURCE_GROUP_ID}" ]]; then
+    echo "Resource_group_id must be passed as second input script argument" >&2
+    exit 1
+fi
+
+# Login to ibmcloud with cli
+attempts=1
+until ibmcloud login -q -r "${REGION}" -g "${RESOURCE_GROUP_ID}" || [ $attempts -ge 3 ]; do
+    attempts=$((attempts+1))
+    echo "Error logging in to IBM Cloud CLI..." >&2
+    sleep 5
+done
+
+# run api-key reset command if apikey for given region + resource group does not already exist
+reset=true
+#key_descriptions=()
+#while IFS='' read -r line; do key_descriptions+=("$line"); done < <(ibmcloud iam api-keys --all --output json | jq -r --arg name "${APIKEY_KEY_NAME}" '.[] | select(.name == $name) | .description')
+#for i in "${key_descriptions[@]}"; do
+#  if [[ "$i" =~ ${REGION} ]] && [[ "$i" =~ ${RESOURCE_GROUP_ID} ]]; then
+#    echo "Found key named ${APIKEY_KEY_NAME} which covers clusters in ${REGION} and resource group ID ${RESOURCE_GROUP_ID}"
+#    reset=false
+#    break
+#  fi
+#done
+key_descriptions=()
+ibmcloud iam api-keys --all --output json | jq -r --arg name "${APIKEY_KEY_NAME}" '.[] | select(.name == $name) | .description' | while IFS='' read -r line; do
+    echo "Line read: $line" # Debug output
+    key_descriptions+=("$line")
+done
+
+# Check if the array is empty
+if [ ${#key_descriptions[@]} -eq 0 ]; then
+    echo "No key descriptions found."
+else
+    # Iterate over key descriptions
+    for i in "${key_descriptions[@]}"; do
+        if [[ "$i" =~ ${REGION} ]] && [[ "$i" =~ ${RESOURCE_GROUP_ID} ]]; then
+            echo "Found key named ${APIKEY_KEY_NAME} which covers clusters in ${REGION} and resource group ID ${RESOURCE_GROUP_ID}"
+            reset=false
+            break
+        fi
+    done
+fi
+
+if [ "${reset}" == true ]; then
+  cmd="ibmcloud ks api-key reset --region ${REGION}"
+  yes | "${cmd}" || echo "Error executing command: ${cmd} && exit $?"
+  # sleep for 10 secs to allow the new key to be replicated across backend DB instances before attempting to create cluster
+  sleep 10
+fi