docker-compose build
docker-compose up -d
docker-compose exec glue bash

$ bash script/preprocess.sh
$ spark-submit example/s3_parquet.py --JOB_NAME=dummy --STAGE=production
$ aws s3 ls s3://test-bucket/spark/ --endpoint-url http://localstack:4566
2021-01-21 14:27:56          0 _SUCCESS
2021-01-21 14:27:55        975 part-00000-9a345736-c9ae-407a-af3c-84ac817d78c9-c000.snappy.parquet
2021-01-21 14:27:56        970 part-00001-9a345736-c9ae-407a-af3c-84ac817d78c9-c000.snappy.parquet

$ aws s3 cp \
    s3://test-bucket/spark/part-00000-9a345736-c9ae-407a-af3c-84ac817d78c9-c000.snappy.parquet \
    foo.snappy.parquet \
    --endpoint-url http://localstack:4566
$ parquet-tools cat foo.snappy.parquet
variable_1 = 1
variable_2 = 1.87
variable_3 = new_york

variable_1 = 4
variable_2 = 2.76
variable_3 = la

variable_1 = 6
variable_2 = 3.3
variable_3 = boston

$ docker-compose exec glue /home/jupyter/jupyter_start.sh

$ mysql -h db -uroot -ptest -P3306
$ psql -h db2 -Upostgres -p5432
$ aws s3 ls --endpoint-url http://localstack:4566

$ mysql -h 127.0.0.1 -uroot -ptest -P3306
$ psql -h 127.0.0.1 -Upostgres -p5432
$ aws s3 ls --endpoint-url http://localhost:4566