Fix factory reset and improve reset_stage CI job

- Add create_default_admin() call to factory reset (admin user wasn't being
  created after reset, only on server restart)
- Add retry logic to reset_stage CI job (3 attempts with 5s delay)
- Use proper context manager for httpx client
- Increase timeout to 120s for reset operation
- Add retry: 1 at job level for transient failures

.gitlab-ci.yml

@@ -154,6 +154,7 @@ reset_stage:
   needs: [integration_test_stage]
   image: deps.global.bsf.tools/docker/python:3.12-slim
   timeout: 5m
+  retry: 1  # Retry once on transient failures
   before_script:
     - pip install --index-url "$PIP_INDEX_URL" httpx
   script:
@@ -162,10 +163,13 @@ reset_stage:
       import httpx
       import sys
       import os
+      import time
 
       BASE_URL = os.environ.get("STAGE_URL", "")
       ADMIN_USER = "admin"
       ADMIN_PASS = "changeme123"  # Default admin password
+      MAX_RETRIES = 3
+      RETRY_DELAY = 5  # seconds
 
       if not BASE_URL:
           print("ERROR: STAGE_URL environment variable not set")
@@ -173,8 +177,8 @@ reset_stage:
 
       print(f"=== Resetting stage environment at {BASE_URL} ===")
 
-      client = httpx.Client(base_url=BASE_URL, timeout=60.0)
+      def do_reset():
-
+          with httpx.Client(base_url=BASE_URL, timeout=120.0) as client:
               # Login as admin
               print("Logging in as admin...")
               login_response = client.post(
@@ -182,8 +186,7 @@ reset_stage:
                   json={"username": ADMIN_USER, "password": ADMIN_PASS},
               )
               if login_response.status_code != 200:
-          print(f"Login failed: {login_response.status_code} - {login_response.text}")
+                  raise Exception(f"Login failed: {login_response.status_code} - {login_response.text}")
-          sys.exit(1)
               print("Login successful")
 
               # Call factory reset endpoint
@@ -195,14 +198,28 @@ reset_stage:
 
               if reset_response.status_code == 200:
                   result = reset_response.json()
-          print(f"Factory reset successful!")
+                  print("Factory reset successful!")
                   print(f"  Database tables dropped: {result['results']['database_tables_dropped']}")
                   print(f"  S3 objects deleted: {result['results']['s3_objects_deleted']}")
                   print(f"  Database reinitialized: {result['results']['database_reinitialized']}")
                   print(f"  Seeded: {result['results']['seeded']}")
-          sys.exit(0)
+                  return True
               else:
-          print(f"Factory reset failed: {reset_response.status_code} - {reset_response.text}")
+                  raise Exception(f"Factory reset failed: {reset_response.status_code} - {reset_response.text}")
+
+      # Retry loop
+      for attempt in range(1, MAX_RETRIES + 1):
+          try:
+              print(f"Attempt {attempt}/{MAX_RETRIES}")
+              if do_reset():
+                  sys.exit(0)
+          except Exception as e:
+              print(f"Attempt {attempt} failed: {e}")
+              if attempt < MAX_RETRIES:
+                  print(f"Retrying in {RETRY_DELAY} seconds...")
+                  time.sleep(RETRY_DELAY)
+              else:
+                  print("All retry attempts failed")
                   sys.exit(1)
       RESET_SCRIPT
   rules:

backend/app/routes.py

@@ -6481,8 +6481,12 @@ def factory_reset(
         # Step 4: Re-seed with default data (need fresh session after schema recreate)
         logger.info("Seeding database with defaults...")
         from .seed import seed_database
+        from .auth import create_default_admin
         fresh_db = SessionLocal()
         try:
+            # Create default admin user first (normally done at startup)
+            create_default_admin(fresh_db)
+            # Then seed other test data
             seed_database(fresh_db)
             fresh_db.commit()
         finally: