[Enhancement] extensive enhance for fullsync

[dataroaring]

Search before asking

• I had searched in the issues and found no similar issues.

Description

Fullsync means sync all data via backup and restore again. It happens when ccr can not handle binlog incrementally, then it starts from beginning again. Internally, fullsync using bakcup with alias and atomic restore to reuse data as mush as possible , thus avoid sync all data again.

So we should let fullsync happen frequently to test it thoughroughly.

Solution

1. add cases for some operations listed by experience as below, test_ds_tbl_truncate is an example.

Resource	Operation	Db sync case	Table sync case	Table sync with alias case	Comment
Database
Table	Create + Drop		-	-
Create + Alter		-	-
Drop + Create		-	-	Schemas are same
Drop + Create		-	-	Schemas are different
Drop		-	-
Truncate	test_ds_tbl_truncate
Rename		-	-
Replace
Alter properties
Alter comment
Partition	Add
	Drop
	Replace
	Modify
	Rename
	Add + Drop
	Drop + Add
	Add + rename
Column	Add key
Drop key
Add value
Drop value
Order by
Rename
Modify comment
Add key + drop
Drop key + add
Add value + drop
Drop value + add
Add column value + modify
Add key column + modify
Rollup	Add
Drop
Rename		-	-
Add column
Drop column
Modify column
Rename column
View	Create
Alter
Drop
Drop + Create
Create + Drop
Index	Add inverted index
Drop inverted index
Add bloom filter
Drop bloom filter
Add NG bloom filter
Drop NG bloom filter
MV	Create
Drop
DML	INSERT INTO OVERWRITE
explicit transaction

2. writing a test tool which generate random operation and ccr syncer take full sync in action randomly.

Are you willing to submit PR?

• Yes I am willing to submit a PR!

Code of Conduct

• I agree to follow this project's Code of Conduct

[dataroaring]

This is a huge issue, please create a new issue for specific task you pick.