HBase Sidecar migration from HDP to CDP

You can use HBase replication and snapshots to migrate HBase tables to the destination cluster.

1. Prepare for the migration. Learn about changes in HBase and what actions you need to take before migrating your data. See Preparing for data migration.
2. Use the HBase shell on the source cluster to take a snapshot of all tables you want to migrate to the destination cluster and then export the snapshots to the destination. See: Using Snapshots (HDP 2) or Using Snapshots in HBase (HDP 3)
3. Export the snapshots to the destination cluster. See: Using Snapshots (HDP 2) Using Snapshots in HBase (HDP 3)
4. Import the snapshot on the destination cluster by running the following command in the HBase shell:

   restore_snapshot <snapshot name> 

5. If you are migrating from HDP 2, you may need to update your client application code. See Known Incompatibilities when Upgrading from CDH and HDP.
6. Review the following topics for details for additional incompatibilities with HDP 2:
   • Remove PREFIX_TREE Data Block Encoding
   • Validate HFiles
   • Check co-processor classes
7. Verify and validate the migrated data on the destination cluster. See Verifying and validating if your data is migrated
8. Test your applications or queries.
9. When you have verified and tested your application code and are ready to move the applications to production, repeat the above steps to synchronize the HBase data with the source cluster.

   You can use the Hash Table and Sync Table tool to synchronize table data. This tool can help improve performance when synchronizing large tables. See Use HashTable and SyncTable Tool.