CollectionMetadata offers a method to get a weak reference to the chunk manager it points to.

It has been observed a case in which this led to the chunk manager being cleared up while still used by an op observer.

The flow that led to hit the error - right after a refineCollectionShardKey was the following:

• The application of an insert oplog entry on a secondary triggered ShardServerOpObserver::onInserts (yes, the observer is also active on secondaries)
• incrementChunkOnInsertOrUpdate got called with a pointer to the old chunk manager as argument, and succeeded to get the old shard key
• Upon calling findIntersectingChunkWithSimpleCollation, the chunk manager got cleared up (because the pointer was not shared) and the chunk info was not found