In an indexer cluster, bucket fixing, or "bucket fixup," is the remedial activity that occurs when a peer node goes offline. The manager node orchestrates the remaining peers in a series of activities, such as replicating buckets and indexing non-searchable bucket copies, with the goal of returning the cluster to a valid and complete state.
Bucket fixing can also occur in a few other circumstances, such as during data rebalancing.
For more information
In Managing Indexers and Clusters of Indexers: