Impala Metadata Replication

Impala metadata replication is performed as an automatic part of Hive replication. Impala replication is only supported between two CDH 5 clusters. The Impala and Hive services must be running on both clusters. As long as these conditions are met, the replication of Impala metadata happens automatically as part of Hive replication.

This ensures that Impala UDFs (user-defined functions) will be available on the target cluster, just as on the source cluster. As part of replicating the UDFs, the binaries in which they're defined are also replicated.