operate-schema set will use --flush as default action

[lance6716]

Feature Request

Is your feature request related to a problem? Please describe:

If we not --flush, which means write a table checkpoint, and following error happens, will cause "drop table in schema tracker because no checkpoint exists" so the operate-schema set is silently disappeared.

Describe the feature you'd like:

as title

Describe alternatives you've considered:

Teachability, Documentation, Adoption, Migration Strategy:

[lance6716]

seems we can't use --flush to create a table structure for fresh new tables 😂

panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x0 pc=0x78cb4ea]

goroutine 2896 [running]:
github.com/pingcap/dm/syncer.(*binlogPoint).MySQLLocation(0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0)
	/Users/lance/Projects/dm/syncer/checkpoint.go:141 +0x6a
github.com/pingcap/dm/syncer.(*RemoteCheckPoint).FlushPointWithTableInfo(0xc0003a6630, 0xc0014118c0, 0xc0023b59a4, 0x4, 0xc0023b59a8, 0x4, 0xc001496a80, 0x0, 0x0)
	/Users/lance/Projects/dm/syncer/checkpoint.go:575 +0x44b
github.com/pingcap/dm/syncer.(*Syncer).OperateSchema(0xc000a58380, 0x902f118, 0xc001d54240, 0xc0015bff20, 0x4, 0xc000a3f408, 0x8, 0xc0023e9498)
	/Users/lance/Projects/dm/syncer/schema.go:93 +0x234e
github.com/pingcap/dm/dm/worker.(*SubTask).OperateSchema(0xc0003ea540, 0x902f118, 0xc001d54240, 0xc0015bff20, 0x4094b5f, 0xc000202480, 0xc0ffffffff, 0xc000000000)
	/Users/lance/Projects/dm/dm/worker/subtask.go:541 +0x289
github.com/pingcap/dm/dm/worker.(*Worker).OperateSchema(0xc000752360, 0x902f118, 0xc001d54240, 0xc0015bff20, 0x0, 0x0, 0x0, 0x0)
	/Users/lance/Projects/dm/dm/worker/worker.go:918 +0x328
github.com/pingcap/dm/dm/worker.(*Server).OperateSchema(0xc000202480, 0x902f118, 0xc001d54240, 0xc0015bff20, 0xc000202480, 0xc001d54240, 0xc0013db918)
	/Users/lance/Projects/dm/dm/worker/server.go:808 +0x81d
github.com/pingcap/dm/dm/pb._Worker_OperateSchema_Handler(0x802c280, 0xc000202480, 0x902f118, 0xc001d54240, 0xc0015bfec0, 0x0, 0x902f118, 0xc001d54240, 0xc0008d7040, 0x93)
	/Users/lance/Projects/dm/dm/pb/dmworker.pb.go:2824 +0x395
google.golang.org/grpc.(*Server).processUnaryRPC(0xc000589680, 0x9055358, 0xc000589e00, 0xc0011b3b00, 0xc00058fad0, 0xabf3f30, 0x0, 0x0, 0x0)
	/Users/lance/Projects/go/pkg/mod/google.golang.org/grpc@v1.27.1/server.go:1024 +0xa23
google.golang.org/grpc.(*Server).handleStream(0xc000589680, 0x9055358, 0xc000589e00, 0xc0011b3b00, 0x0)
	/Users/lance/Projects/go/pkg/mod/google.golang.org/grpc@v1.27.1/server.go:1313 +0x13b8
google.golang.org/grpc.(*Server).serveStreams.func1.1(0xc00055d0a0, 0xc000589680, 0x9055358, 0xc000589e00, 0xc0011b3b00)
	/Users/lance/Projects/go/pkg/mod/google.golang.org/grpc@v1.27.1/server.go:722 +0xe7
created by google.golang.org/grpc.(*Server).serveStreams.func1
	/Users/lance/Projects/go/pkg/mod/google.golang.org/grpc@v1.27.1/server.go:720 +0xb9

cc @GMHDBJD How could we set checkpoint to avoid "drop table in schema tracker because no checkpoint exists"?