# Dumpling Snapshot

+ Confirm the timestamp of the target data recover

+ dumpling backup target data history
  + `dumpling -h 172.xx.xx.xx -P 4000 -uroot -p xxx -t 32 -F 64MiB -B target_db --snapshot "yyyy-mm-dd hh:mi:ss" -o /tmp/dumpling_dir` 

+ TIDB-Lighting import data

------------------------------------------------------------

Dumpling supports consistent export of data, and TSO to specify dump of TiDB data at a snapshot point in time. If the GC is met, data recover can be implement using dumpling. The key parameters are as follows:
 —snapshot snapshot tso, only takes effect when consistency=snapshot

The data recover reference process is as follows:
1. Confirm the timestamp of the target data recover

2. Dumpling backup target data history
dumpling -h 172.xx.xx.xx -P 4000 -uroot -p xxx -t 32 -F 64MiB -B target_db —snapshot “yyyy-mm-dd hh:mi:ss” -o /tmp/dumpling_dir

3. TIDB-Lighting import data
1) Import the original cluster remap_table, and then perform subsequent data recover processing, such as generating reverse SQL
2) After importing the original name of the temporary cluster, perform subsequent data recover processing, such as generating reverse SQL

Unlike DDL's Flashback, Recover Table, this method requires secondary processing of the recovered data, such as generating reverse SQL, and then applying it to the target business table.