Currently, the shard limits the number of threads refreshing the config metadata here. This limit is very crude and does not differentiate refresh for different namespaces. It is possible that the all the ticket holders are trying to refresh the same collection. In this case, it limits parallelism while doing wasted work.
- is depended on by
-
SERVER-20854 Don't refresh database and collection information from config server while holding a mutex
- Closed
- is duplicated by
-
SERVER-28386 Skip invalidated collections in CatalogCache::onStaleConfigError
- Closed
-
SERVER-27475 mongos should request an update only of the collection not found
- Closed