PolarDB-X 标准版:16C64G(1台)
RDS MySQL5.7: 4C8G(4台)
直连RDS:
MySQL [hlj_order]> SELECT count(*) FROM `order` AS t1 LEFT JOIN `order_item` AS t2 ON t1.`order_seq` = t2.`order_seq` LEFT JOIN `order_item_price` AS t3 ON t3.`item_seq` = t2.`item_seq` GROUP BY t1.`user_id`; 。。。。。。。。。。。。。 | 1 | | 1 | | 2 | | 2 | | 4 | | 2 | | 1 | | 1 | | 3 | | 9 | +----------+ 4599756 rows in set (5 min 55.96 sec)
三张表关联查询,耗时5min55s。
查看监控CPU和IO消耗:


PolarDB-X +RDS x4分库分表:
MySQL [polardb2]> SELECT count(*) FROM `order` AS t1 LEFT JOIN `order_item` AS t2 ON t1.`order_seq` = t2.`order_seq` LEFT JOIN `order_item_price` AS t3 ON t3.`item_seq` = t2.`item_seq` GROUP BY t1.`user_id`; 。。。。。。。。。。。。。。。。 | 1 | | 1 | | 1 | | 1 | | 1 | | 1 | | 1 | | 1 | | 1 | | 1 | | 1 | | 1 | +----------+ 4599756 rows in set (1 min 26.34 sec)
耗时1min26s。
查看监控CPU和IO资源消耗:


总结:对于复杂查询,PolarDB-X +RDS分库分表的架构,在性能和资源消耗方面还是有一定的优势。
来自 “ ITPUB博客 ” ,链接:https://blog.itpub.net/30135314/viewspace-2751111/ ,如需转载,请注明出处,否则将追究法律责任。