Definition
flushRouterConfigflushRouterConfigmarks the cached routing table for a sharded cluster as stale, causing the next operation that requests the routing table to refresh the cache. Management of the routing table cache is generally handled automatically by the cluster. You should only need to run theflushRouterConfigcommand manually in specific instances, as indicated under Considerations.
Note
Starting in MongoDB 4.4, running flushRouterConfig
is no longer required after executing the movePrimary or
dropDatabase commands. These two commands now
automatically refresh a sharded cluster's routing table as needed
when run.
Syntax
Starting in MongoDB 4.0.6 (and 3.6.11), the
flushRouterConfig is available on both
mongos and mongod instances, and has the
following syntax:
Flush the cache for a specified collection when passed in a collection namespace parameter:
db.adminCommand({ flushRouterConfig: "<db.collection>" } ) Flush the cache for a specified database and all of its collections when passed in a database namespace parameter:
db.adminCommand({ flushRouterConfig: "<db>" } ) Flush the cache for all databases and their collections when run without a parameter or passed in a non-string scalar value (e.g.
1):db.adminCommand("flushRouterConfig") db.adminCommand( { flushRouterConfig: 1 } )
Note
In MongoDB 4.0.5 and earlier (and 3.6.10 and earlier),
flushRouterConfig is only available for
mongos instances and can only flush the cache for all
databases and their collections:
db.adminCommand("flushRouterConfig") db.adminCommand( { flushRouterConfig: 1 } )
Considerations
Generally, a sharded cluster's routing table is automatically refreshed
as needed as part of normal operation. However, you should manually
issue the flushRouterConfig command in the following
instances:
If running MongoDB 4.2 or earlier, after executing the
movePrimaryordropDatabasecommands.If running MongoDB 4.2.2 or earlier (or 4.0.14 or earlier), after manually clearing the
jumboflag from a chunk that could not be split (i.e. an indivisible chunk). See ClearjumboFlag for more information.Before running the
db.collection.getShardDistribution()command. This ensures that the command returns the most up-to-date information.