[new feature] support synchronizing multiple tables

[zhangjinpeng87]

Background

Currently the tidb2dw tool is single process model, each process just support single table synchronizing pipeline. If upstream has hundreds of tables, the tidb2dw tool can not easily handle it. There is a lot of maintenance cost, as well as memory and other resource cost.

Proposal

• tidb2dw supports a set of API to add/drop/pause/resume synchronizing jobs
• tidb2dw supports running multiple synchronizing jobs

[Lloyd-Pottiger]

supports running multiple replicate tables by #58