To reduce the pressure on hbase clusters and improve performance, we move data from the HBase library to another store, which records the process of data recovery when I need to calculate data
Directory:
- Azure Storage Explorer Tools
- Data replication
- Metadata Repair
- Test validation
Azure Storage Explorer Tools
- Free tools: http://storageexplorer.com/
- After you install locally, connect to azure storage, such as:
- Description
- Replication performance is better when the source and destination stores belong to the same region, and note that I'm using a low-storage explorer where there is no progress bar at copy
- In the target hbase cluster, the repair command automatically creates a table when the complete directory is copied past, using hbase hbck to repair the command if the table does not exist in the target cluster
Data replication
- Description: The full directory copy, using the tool to copy, paste, copy of the performance can also
- The bms:16:30-16:40 takes 10 minutes and the data size is as follows:
- Power:17:10-17:47 takes 37 minutes, data volume size:
- Folder Rename: The tool will create a new folder, copy the contents of the source file to the new folder, and then delete the source folder, the amount of data, the more laborious
Metadata Repair
- After the data is copied, enter the HBase installation directory for metadata repair, quick fix command:hbase hbck-repair-ignoreprecheckpermission
- After successful repair, such as:
Test validation
- Enter HBase shell Command mode
- Command: List (for example, Description table fix OK)
- Command: Scan ' Terminaldata ', {limit=>1} (for example, read and write OK)
Azure BLOB Storage T-level hbase table recovery