##先新建复制表库名、复制表
create databases if not exists schema_name_test on cluster default_cluster;
create table if not exists schema_name_test.tableName on cluster default_cluster (`id` String, `name` String, `createTime` DateTime) engine=ReplicatedMergeTree('/clickhouse/tables/{shard}/schema_name_test.tableName','{replica}') partition by toYYYYMM(createTime) order by id;
#先删除再创建报错
drop table schema_name_test.tableName on cluster default_cluster;
create table if not exists schema_name_test.tableName on cluster default_cluster (`id` String, `name` String, `createTime` DateTime) engine=ReplicatedMergeTree('/clickhouse/tables/{shard}/schema_name_test.tableName','{replica}') partition by toYYYYMM(createTime) order by id;
错误描述
DB::Exception: Replica /clickhouse/tables/01-03/xxxxxx/xxx/replicas/xxx already exists…
原因因为复制表需要在zookeeper上建立一个路径，存放相关数据。clickhouse默认的库引擎是原子数据库引擎，删除Atomic数据库中的表后，它不会立即删除，而是会在480秒后删除。由下面这个参数控制：
config.xml
解决方法480
1.修改config.xml的配置database_atomic_delay_before_drop_table_sec为0
2.最原始方法，等待480秒后再重新建表
为了保险起见，楼主用的第二种方法，再次创建成功，不建议修改配置
