Overlap validation and repair tool
Less than 1 minute
Overlap validation and repair tool
The Overlap Validation And Repair tool is used to validate the resource files in sequence space, and repair overlaps.
The validation function can be run in any scenario. Confirmation is required after overlapping files are found. Typing 'y' will perform the repair.
The repair function must be run when corresponding DataNode is stopped and there are no unfinished compaction task in all data dirs.
To make sure there are no unfinished compaction tasks, you can modify the config files to set enable compaction items to false, and restart DataNode waiting compaction recover task to finish.
Then stop the DataNode and run this tool.
Usage
#MacOs or Linux
./check-overlap-sequence-files-and-repair.sh [sequence_data_dir1] [sequence_data_dir2]...
# Windows
.\check-overlap-sequence-files-and-repair.bat [sequence_data_dir1] [sequence_data_dir2]...
Example
./check-overlap-sequence-files-and-repair.sh /data1/sequence/ /data2/sequence
This example validate two data dirs: /data1/sequence/, /data2/sequence.