Revert "NullPool for the celery worker." (#1488)

apache · Oct 31, 2016 · 4dc959a · 4dc959a
1 parent 518fbf5
commit 4dc959a
Showing 1 changed file with 2 additions and 9 deletions.
diff --git a/caravel/sql_lab.py b/caravel/sql_lab.py
@@ -10,7 +10,6 @@
  app, db, models, utils, dataframe, results_backend)
 from caravel.db_engine_specs import LimitMethod
 from caravel.jinja_context import process_template
-from sqlalchemy.pool import NullPool
 QueryStatus = models.QueryStatus
 
 celery_app = celery.Celery(config_source=app.config.get('CELERY_CONFIG'))
@@ -47,15 +46,9 @@ def create_table_as(sql, table_name, schema=None, override=False):
  return exec_sql.format(**locals())
 
 
-@celery_app.task(bind=True)
-def get_sql_results(self, query_id, return_results=True, store_results=False):
+@celery_app.task
+def get_sql_results(query_id, return_results=True, store_results=False):
  """Executes the sql query returns the results."""
-
- # disable pooling for the celery task to prevent
- # MySQL server gone issue.
- if not self.request.called_directly:
- db.engine.pool = NullPool(db.engine.create)
-
  session = db.session()
  session.commit() # HACK
  query = session.query(models.Query).filter_by(id=query_id).one()