Check Kudu Cluster Health with ksck

Kudu provides the ksck tool to check cluster health and gather detailed information. It can detect issues such as under-replicated tablets, unreachable tablet servers, or tablets without a leader.

Run ksck to Check the Cluster Health

  • Run the command as the kudu user.
  • To see all available commands and options, use:
Bash
Copy
  • If the cluster is healthy, ksck prints cluster information, a success message, and returns 0.
  • If there are errors, ksck returns a non-zero status code.

Example: ksck Output

If an error occurs—for example, when a tablet server is down—ksck returns a non-zero status code and displays output similar to the following:

Bash
Copy

Verify Cluster Data Consistency Using ksck

You can use the --checksum_scan option to verify data consistency across the cluster. This option scans the tablets and compares their data.

  • Use --tables to limit the scan to specific tables.
  • Use --tablets to limit the scan to specific tablets.

Note The --tablets option refers to tablet IDs, not tablet servers.

To retrieve tablet IDs, run:

Bash
Copy

The first column after the T is the ID of each tablet.

For more details, see Apache Documentation.

Type to search, ESC to discard
Type to search, ESC to discard
Type to search, ESC to discard
  Last updated