[Improvement] Keep encoded size of block consistent between client and server

### Code of Conduct

- [X] I agree to follow this project's [Code of Conduct](https://www.apache.org/foundation/policies/conduct)

### Search before asking

- [X] I have searched in the [issues](https://github.com/apache/incubator-uniffle/issues?q=is%3Aissue) and found no similar issues.

### What would you like to be improved?

Keep the client side `ShuffleBlockInfo#getSize` and server side `ShufflePartitionedBlock#getSize` with the consistent `encoded header length`.

### How should we improve?

Client send the `encoded header length` to the server, then server use the received `encoded header length` for compute the `getSize` method.

### Are you willing to submit PR?

- [X] Yes I am willing to submit a PR!

Start by locating ShuffleBlockInfo#getSize and ShufflePartitionedBlock#getSize, then trace how the client encodes and sends the block header and how the server receives it. Done means the client-provided encoded header length reaches the server and both getSize methods compute consistently with it.

Written by the indexing model from the issue text.