Config

These environment variables can be modified and exported before running cortex-installer.sh commands. Alternatively, a configuration file may be provided to cortex-installer.sh via the --config flag (e.g. cortex-installer.sh --config=./config.sh install operator). Default values are shown.

# config.sh
# AWS credentials
export AWS_ACCESS_KEY_ID="***"
export AWS_SECRET_ACCESS_KEY="***"
# The name of the CloudWatch log group Cortex will use
export CORTEX_LOG_GROUP="cortex"
# The name of the S3 bucket Cortex will use
export CORTEX_BUCKET="cortex-[RANDOM_ID]"
# The AWS region Cortex will use
export CORTEX_REGION="us-west-2"
# The name of the Kubernetes namespace Cortex will use
export CORTEX_NAMESPACE="cortex"
# Flag to enable collecting error reports and usage stats. If flag is not set to either "true" or "false", you will be prompted.
export CORTEX_ENABLE_TELEMETRY=""
# Image paths
export CORTEX_IMAGE_ARGO_CONTROLLER="cortexlabs/argo-controller:0.3.0"
export CORTEX_IMAGE_ARGO_EXECUTOR="cortexlabs/argo-executor:0.3.0"
export CORTEX_IMAGE_FLUENTD="cortexlabs/fluentd:0.3.0"
export CORTEX_IMAGE_NGINX_BACKEND="cortexlabs/nginx-backend:0.3.0"
export CORTEX_IMAGE_NGINX_CONTROLLER="cortexlabs/nginx-controller:0.3.0"
export CORTEX_IMAGE_OPERATOR="cortexlabs/operator:0.3.0"
export CORTEX_IMAGE_SPARK="cortexlabs/spark:0.3.0"
export CORTEX_IMAGE_SPARK_OPERATOR="cortexlabs/spark-operator:0.3.0"
export CORTEX_IMAGE_TF_SERVE="cortexlabs/tf-serve:0.3.0"
export CORTEX_IMAGE_TF_TRAIN="cortexlabs/tf-train:0.3.0"
export CORTEX_IMAGE_TF_API="cortexlabs/tf-api:0.3.0"
export CORTEX_IMAGE_TF_TRAIN_GPU="cortexlabs/tf-train-gpu:0.3.0"
export CORTEX_IMAGE_TF_SERVE_GPU="cortexlabs/tf-serve-gpu:0.3.0"
export CORTEX_IMAGE_PYTHON_PACKAGER="cortexlabs/python-packager:0.3.0"