-
Type: Bug
-
Resolution: Fixed
-
Priority: Major - P3
-
Affects Version/s: None
-
Component/s: Testing Infrastructure
-
None
-
Fully Compatible
-
ALL
-
v4.0, v3.6
-
TIG 2018-07-02, TIG 2018-07-16
The logic in the FSMWorkloadTestCase class is faulty because it sets TestData.sameDB=true when same_db=false and TestData.sameCollection=true when same_collection=false. However, these options currently have no effect because resmoke_runner.js doesn't define clusterOptions.sameDB and clusterOptions.sameCollection based on their associated TestData equivalents.
# We use a global incrementing counter as a prefix for the database name to avoid any # collection lifecycle related issues in sharded clusters. This more closely matches how # uniqueDBName() and uniqueCollName() would have returned distinct values when called once # for each FSM workload in the entire schedule by runner.js. test_prefix = self.db_name_prefix if self.db_name_prefix else "test" test_data["dbNamePrefix"] = "{}{:d}_".format(test_prefix, count) if not self.same_db: test_data["sameDB"] = True if not self.same_collection: test_data["sameCollection"] = True
- is caused by
-
SERVER-30204 Create resmoke.py hook that drops all databases without restarting the cluster
- Closed