From 04411d82228ce0b9cbd16f78c008dbed99cb8f1c Mon Sep 17 00:00:00 2001
From: Oleksandr Mazur <mazur.ol.sr@gmail.com>
Date: Fri, 14 Nov 2025 09:40:09 +0200
Subject: [PATCH] moved embucket start into run.sh

---
 .github/workflows/dbt-gitlab-run.yml          |  9 ----
 test/dbt_integration_tests/dbt-gitlab/run.sh  | 22 ++++----
 .../dbt-gitlab/upload.py                      | 51 -------------------
 3 files changed, 13 insertions(+), 69 deletions(-)
 delete mode 100644 test/dbt_integration_tests/dbt-gitlab/upload.py

diff --git a/.github/workflows/dbt-gitlab-run.yml b/.github/workflows/dbt-gitlab-run.yml
index 2a80f6e61..9d9ed12fa 100644
--- a/.github/workflows/dbt-gitlab-run.yml
+++ b/.github/workflows/dbt-gitlab-run.yml
@@ -21,15 +21,6 @@ jobs:
         with:
           submodules: true
 
-      - name: Start Embucket Server (Docker)
-        run: |
-          docker run -d \
-            --name embucket-server \
-            -p 3000:3000 \
-            embucket/embucket
-          echo "Starting Embucket server..."
-          sleep 20
-
       - name: Set up Python
         uses: actions/setup-python@v6
         with:
diff --git a/test/dbt_integration_tests/dbt-gitlab/run.sh b/test/dbt_integration_tests/dbt-gitlab/run.sh
index cbfd2467f..9285fea06 100755
--- a/test/dbt_integration_tests/dbt-gitlab/run.sh
+++ b/test/dbt_integration_tests/dbt-gitlab/run.sh
@@ -1,5 +1,18 @@
 #!/bin/bash
 
+# Start Embucket container with NO persistent storage
+echo "Starting Embucket container with clean environment..."
+docker run -d --rm --name em \
+  -p 3000:3000 \
+  -p 8080:8080 \
+  --env OBJECT_STORE_BACKEND=memory \
+  --env SLATEDB_PREFIX=memory \
+  --env DATA_FORMAT=arrow \
+  embucket/embucket-labs >/dev/null 2>&1
+
+echo "✓ Embucket container started successfully with CLEAN environment!"
+echo ""
+
 # Parse --target and --model arguments
 while [[ "$#" -gt 0 ]]; do
   case $1 in
@@ -89,15 +102,6 @@ echo "Installing the requirements"
 pip install -r requirements.txt >/dev/null 2>&1
 echo ""
 
-# Load data and create embucket catalog if the embucket is a target 
-echo "###############################"
-echo ""
-echo "Creating embucket database"
-if [ "$DBT_TARGET" = "embucket" ]; then
-  $PYTHON_CMD upload.py
-fi
-echo ""
-
 mkdir -p assets
 
 # Run DBT commands
diff --git a/test/dbt_integration_tests/dbt-gitlab/upload.py b/test/dbt_integration_tests/dbt-gitlab/upload.py
deleted file mode 100644
index 053efd515..000000000
--- a/test/dbt_integration_tests/dbt-gitlab/upload.py
+++ /dev/null
@@ -1,51 +0,0 @@
-import os
-
-import snowflake.connector
-
-url = "http://localhost:3000"
-database = "embucket"
-schema = "public"
-
-
-def bootstrap(table_name="sample_table"):
-    cursor = get_cursor()
-
-    # Volume
-    cursor.execute(f"CREATE EXTERNAL VOLUME IF NOT EXISTS test STORAGE_LOCATIONS = (\
-        (NAME = 'file_vol' STORAGE_PROVIDER = 'FILE' STORAGE_BASE_URL = '{os.getcwd()}/data'))")
-    print(f"Volume 'test' created at '{os.getcwd()}/data' created or already exists.")
-
-    # Database
-    cursor.execute(f"CREATE DATABASE IF NOT EXISTS {database} EXTERNAL_VOLUME = test")
-    print(f"Database {database} created or already exists.")
-
-    # Schema
-    cursor.execute(f"CREATE SCHEMA IF NOT EXISTS {database}.{schema}")
-    print(f"Schema {database}.{schema} created or already exists.")
-
-    # Sample Table
-    cursor.execute(f"CREATE TABLE IF NOT EXISTS {database}.{schema}.{table_name} \
-        (id INT, name VARCHAR, created_at TIMESTAMP)")
-    print(f"Sample table {database}.{schema}.{table_name} created or already exists.")
-
-
-def get_cursor():
-    con = snowflake.connector.connect(
-        host=os.getenv("EMBUCKET_HOST", "localhost"),
-        port=os.getenv("EMBUCKET_PORT", 3000),
-        protocol=os.getenv("EMBUCKET_PROTOCOL", "http"),
-        user=os.getenv("EMBUCKET_USER", "embucket"),
-        password=os.getenv("EMBUCKET_PASSWORD", "embucket"),
-        account=os.getenv("EMBUCKET_ACCOUNT", "acc"),
-        warehouse=os.getenv("EMBUCKET_WAREHOUSE", ""),
-        database=os.getenv("EMBUCKET_DATABASE", database),
-        schema=os.getenv("EMBUCKET_SCHEMA", schema),
-        session_parameters={
-            "QUERY_TAG": "dbt-testing",
-        },
-    )
-    return con.cursor()
-
-
-if __name__ == "__main__":
-    bootstrap()