Enhancement

As title, sometimes, after a PiTR, we may want to do checksum by the native ADMIN CHECKSUM statement, but that will almost impossible to get CRC64XOR right after a table ID rewriting.

We may provide a subcommand for BR, that generates rewrite rule by a specified backup storage and runs checksum in the target cluster.

An example:

> br operator checksum -s $(get-backup) -f 'test.*' -u tc-pd:2379 --checksum-concurrency 256 --table-concurrency 20

[{"db_name":"test","table_name":"sbtest1","checksum":14313396040806895863,"total_bytes":535765921376,"total_kvs":1319620496}]