Initial sync does not maintain pinned history. So if the ReshardingCollectionCloner or ReshardingTxnCloner aggregation targets a node that went through initial sync, it could fail with a SnapshotUnavailable error. Similarly, if the ReshardingOplogFetcher aggregation targets a node that went through initial sync, it could fail with a OplogQueryMinTsMissing error. Please see BF-39741. SERVER-107404 will add internal retries for this. However, it cannot be done yet due to the reason described here. So for now we should add an override to retry resharding on snapshot related errors due to initial sync.