.envrc-gcp

#!/usr/bin/env bash
#  vim:ts=4:sts=4:sw=4:et
#
#  Author: Hari Sekhon
#  Date: Mon Feb 22 17:42:01 2021 +0000
#
#  https://github.com/HariSekhon/DevOps-Bash-tools
#
#  License: see accompanying Hari Sekhon LICENSE file
#
#  If you're using my code you're welcome to connect with me on LinkedIn and optionally send me feedback to help steer this or other code I publish
#
#  https://www.linkedin.com/in/HariSekhon
#

# ============================================================================ #
#                              G C P   D i r E n v
# ============================================================================ #

# https://direnv.net/man/direnv-stdlib.1.html

# See Also:
#
#   .envrc
#   .envrc-aws
#   .envrc-kubernetes

# direnv stdlib - loads .envrc from parent dir up to /
#
# useful to accumulate parent and child directory .envrc settings eg. adding Kubernetes namespace, ArgoCD app etc.
#
# bypasses security authorization though - use with care
#source_up
#
# source_up must be loaded before set -u otherwise gets this error:
#
#   direnv: loading .envrc
#   /bin/bash: line 226: $1: unbound variable

set -euo pipefail
[ -n "${DEBUG:-}" ] && set -x
srcdir="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"

if [ -n "${CI:-}" ]; then
    exit 0
fi

# https://cloud.google.com/sdk/gcloud/reference/config

# If using other services, infer the environment variables to put below by reading:
#
#   gcloud topic configurations
#       or
#   gcloud config set --help

gcloud_config(){
    local config="${1:-}"
    if [ -z "$config" ]; then
        echo "no config passed to gcloud_config() function" >&2
        return 1
    fi
    if [ -z "${CI:-}" ]; then
        return
    fi
    # don't waste time if not using GCloud SDK, ie. not found in $PATH
    if type -P gcloud; then
        # protect from setting this if the config does exist as this can cause auth problems by unsetting the core.account
        if gcloud config configurations list --format='get(name)' | grep -q "^$config$"; then
            export CLOUDSDK_ACTIVE_CONFIG_NAME="$config"
        fi
    fi
}

#gcloud_config dev
#gcloud_config staging
#gcloud_config production

# XXX: Edit
export CLOUDSDK_CORE_PROJECT=myproject

echo "CLOUDSDK_CORE_PROJECT=$CLOUDSDK_CORE_PROJECT"
echo

# XXX: Edit
export CLOUDSDK_COMPUTE_REGION="${CLOUDSDK_COMPUTE_REGION:-eu-west-2}"

echo "CLOUDSDK_COMPUTE_REGION=$CLOUDSDK_COMPUTE_REGION"
echo

REGION="$CLOUDSDK_COMPUTE_REGION"

# you should probably not set CLOUDSDK_COMPUTE_ZONE
#
# 'gcloud compute ssh' will auto-determine the zone
#
# setting CLOUDSDK_COMPUTE_ZONE explicitly breaks the above command in 2/3 cases due to a VM being in a different zone:
#
#    ERROR: (gcloud.compute.ssh) Could not fetch resource:
#     - The resource 'projects/<MY_PROJECT>/zones/<ZONE>/instances/<VM_NAME>' was not found
#
# gcp/gce_ssh.sh script in this repo can work around that if you do set this
#
#export CLOUDSDK_COMPUTE_ZONE="${REGION}-a" # or b or c

export CLOUDSDK_AI_REGION="$REGION"
export CLOUDSDK_AI_PLATFORM_REGION="$REGION"
export CLOUDSDK_DATAPROC_REGION="$REGION"
export CLOUDSDK_DEPLOY_REGION="$REGION"
export CLOUDSDK_FILESTORE_REGION="$REGION"
export CLOUDSDK_FUNCTIONS_REGION="$REGION"
export CLOUDSDK_MEMCACHE_REGION="$REGION"
export CLOUDSDK_REDIS_REGION="$REGION"
export CLOUDSDK_RUN_REGION="$REGION"
export CLOUDSDK_RUN_CLUSTER_LOCATION="$REGION"
export CLOUDSDK_VMWARE_REGION="$REGION"

# XXX: Edit
export CLOUDSDK_RUN_PLATFORM=managed
#export CLOUDSDK_RUN_PLATFORM=gke
#export CLOUDSDK_RUN_PLATFORM=kubernetes
#export CLOUDSDK_RUN_CLUSTER=mycluster

export CLOUDSDK_GCLOUDIGNORE_ENABLED=True
#export CLOUDSDK_BUILDS_USE_KANIKO=True

# XXX: Edit, or remove if only have 1 cluster in project, will auto-determine below
export CLOUDSDK_CONTAINER_CLUSTER=mycluster  # GKE cluster name

# safer but slower
#gke_clusters=()
#while IFS='' read -r line; do
#    gke_clusters+=("$line")
#done < <(gcloud container clusters list --format='get(name)')
#if [ "${#gke_clusters[@]}" -eq 1 ]; then
#    export CLOUDSDK_CONTAINER_CLUSTER="${gke_clusters[*]}"
#fi

gke_clusters="$(
    gcloud container clusters list --format='get(name)' |
    sed '/^[[:space:]]*$/d'
)"

if [ -n "$gke_clusters" ]; then
    num_gke_clusters="$(grep -c . <<< "$gke_clusters")"
    echo "GKE Clusters ($num_gke_clusters):"
    echo
    echo "$gke_clusters"
    echo
    # If GKE_CLUSTER isn't set and there is only one GKE cluster in this account and region, then use it
    if [ -z "${GKE_CLUSTER:-}" ]; then
        if [ "$num_gke_clusters" = 1 ]; then
            GKE_CLUSTER="$gke_clusters"
        fi
    fi
else
    num_gke_clusters=0
fi

# alternatively call gke_kube_context() function in .envrc-kubernetes which will do this
# and comment out auto-running kube_context() on sourcing .envrc-kubernetes
if [ -n "${CLOUDSDK_CONTAINER_CLUSTER:-}" ]; then
    echo "CLOUDSDK_CONTAINER_CLUSTER=$CLOUDSDK_CONTAINER_CLUSTER"
    echo
    # kubectl context is easily created by running adjacent aws_kube_creds.sh script first

    export GKE_CONTEXT="gke_${CLOUDSDK_CORE_PROJECT}_${CLOUDSDK_COMPUTE_REGION}_${CLOUDSDK_CONTAINER_CLUSTER}"

    # XXX: safer to inline .envrc-kubernetes if you're worried about changes to it bypassing 'direnv allow' authorization
    # shellcheck disable=SC1090,SC1091
    . "$srcdir/.envrc-kubernetes" "$GKE_CONTEXT" ${GKE_NAMESPACE:+"$GKE_NAMESPACE"}
fi

# pull the secret using this command whenever you need it:
#
#   gcp_secret_get.sh "$JENKINS_ADMIN_PASSWORD_GCP_SECRET" | copy_to_clipboard.sh
#
export JENKINS_ADMIN_PASSWORD_GCP_SECRET="jenkins-admin-password"