Potential Resource Leaks and Tenant Isolation Issues in Changefeed Creation and SchemaStorage Initialization

1. When creating a changefeed, RegisterKeyspace is directly called during API request processing with context.Background(). If the subsequent changefeed creation process fails, this will lead to resource leaks.

	// The ctx's lifecycle is the same as the HTTP request.
	// The schema store may use the context to fetch database information asynchronously.
	// Therefore, we cannot use the context of the HTTP request.
	// We create a new context here.
	schemaCxt := context.Background()
	if err := schemaStore.RegisterKeyspace(schemaCxt, keyspaceName); err != nil {
		_ = c.Error(err)
		return
	}

2. There are numerous fatal and panic logs in the schemaStorage initialization path. After supporting multi-keyspace, these logs may become a hidden danger - if one tenant encounters problems, it will affect all tenants.

if err := os.RemoveAll(dbPath); err != nil {
		**log.Panic("fail to remove path")**
	}

	isDataReusable := false
	if exists(dbPath) {
		isDataReusable = true
		db := openDB(dbPath)
		// check whether the data on disk is reusable
		gcTs, err := readGcTs(db)
		if err != nil {
			isDataReusable = false
		}
		if gcSafePoint < gcTs {
			**log.Panic("gc safe point should never go back")**
		}