ShardServerCatalogCacheLoader::waitForCollectionFlush calls waitForActiveTaskCompletion which does an uninterruptible wait. This can lead to deadlocks when stepping down.
The same pattern exists for ShardServerCatalogCacheLoader::waitForDatabaseFlush.
We should make these waits interruptible
- is caused by
-
SERVER-56648 Make dropCollection resilient to network partitions
-
- Closed
-
- related to
-
SERVER-60161 Deadlock between config server stepdown and _configsvrRenameCollectionMetadata command
-
- Closed
-