-
Type: Task
-
Resolution: Fixed
-
Priority: Major - P3
-
Affects Version/s: None
-
Component/s: None
-
Fully Compatible
-
Execution Team 2021-11-15, Execution Team 2021-11-29, Execution Team 2021-12-13, Execution Team 2021-12-27
-
15
createIndex( { key: <clusterKey>, 'clustered': true}) should behave as follows
(1) no-op if the collection is already clustered by 'cluster key'
(2) implicitly create the collection on clusterKey provided the collection does not yet exist.
(3) Throw if the collection is already clustered, but not on the clusterKey provided in the command
Old Description:
There are several places in the code where listIndexes results are assumed to be standard indexes (non-clustered).
Since the 'clustered' index for a clustered collection is also outputted in listIndexes, we want to prevent trying to build it / handle it as a standard index.
eg) SERVER-60979
- is depended on by
-
SERVER-60123 Support using clusterKey {_id : 1} for shardKey in a sharded collection
- Closed
-
SERVER-60979 TenantCollectionCloner::listIndexesStage() should filter out clustered indexes
- Closed
- related to
-
SERVER-62055 Enable tests in clustered_collection_passthrough that try to recreate the clustered index
- Closed