fetch data from tikv every time after resuming the changefeed even though subscribe the same table

Run the following steps:

1. deploy one ticdc node.
2. create 4 changefeeds, all subscribe all tables, no filter set.
3. pause the changefeed
4. run the workload, write in a lot of data to trigger incremental scan
5. resume changefeed one by one.

From images above we can know that after each changefeed resumed, all triggers incremental scan tasks.

Since all tables share the same, those data should can be shared, to reduce the cost on fetch data from tikv and write into the local disk storage.

Only fetch data after resume the first changefeed is expected