天天看点

PostgreSQL 并行计算解说 之18 - parallel merge join

标签

PostgreSQL , cpu 并行 , smp 并行 , 并行计算 , gpu 并行 , 并行过程支持

https://github.com/digoal/blog/blob/master/201903/20190317_10.md#%E8%83%8C%E6%99%AF 背景

PostgreSQL 11 优化器已经支持了非常多场合的并行。简单估计,已支持27余种场景的并行计算。

parallel seq scan                                      
                                      
parallel index scan                                      
                                      
parallel index only scan                                      
                                      
parallel bitmap scan                                      
                                      
parallel filter                                      
                                  
parallel hash agg                                  
                                  
parallel group agg                                  
                                      
parallel cte                                      
                                      
parallel subquery                                      
                                      
parallel create table                                      
                                      
parallel create index                                      
                                      
parallel select into                                      
                                      
parallel CREATE MATERIALIZED VIEW                                      
                                      
parallel 排序 : gather merge                                       
                                      
parallel nestloop join                                      
                                      
parallel hash join                                      
                                      
parallel merge join                                      
                                      
parallel 自定义并行聚合                                      
                                      
parallel 自定义并行UDF                                      
                                      
parallel append                                      
                                      
parallel union                                      
                                      
parallel fdw table scan                                      
                                      
parallel partition join                                      
                                      
parallel partition agg                                      
                                      
parallel gather                              
                      
parallel gather merge                      
                                      
parallel rc 并行                                      
                                      
parallel rr 并行                                      
                                      
parallel GPU 并行                                      
                                      
parallel unlogged table                                       
           

接下来进行一一介绍。

关键知识请先自行了解:

1、优化器自动并行度算法 CBO

《PostgreSQL 9.6 并行计算 优化器算法浅析》 《PostgreSQL 11 并行计算算法,参数,强制并行度设置》

https://github.com/digoal/blog/blob/master/201903/20190317_10.md#parallel-merge-join parallel merge join

并行merge JOIN

数据量:10亿 join 10亿 on (i=i)。

场景 数据量 关闭并行 开启并行 并行度 开启并行性能提升倍数
10亿 join 10亿 using (i) where t1.i<10000000 3.2 秒 1 秒 8 3.2 倍

https://github.com/digoal/blog/blob/master/201903/20190317_10.md#1%E5%85%B3%E9%97%AD%E5%B9%B6%E8%A1%8C%E8%80%97%E6%97%B6-32-%E7%A7%92 1、关闭并行,耗时: 3.2 秒。

postgres=# explain select count(*) from table5 t1 join table5 t2 using (i) where t1.i<10000000 ;     
                                                  QUERY PLAN                                                     
---------------------------------------------------------------------------------------------------------------  
 Aggregate  (cost=20815513.37..20815513.38 rows=1 width=8)  
   ->  Merge Join  (cost=1.15..20794436.78 rows=8430637 width=0)  
         Merge Cond: (t1.i = t2.i)  
         ->  Index Only Scan using idx_table5_2 on table5 t1  (cost=0.57..172964.32 rows=8430637 width=4)  
               Index Cond: (i < 10000000)  
         ->  Index Only Scan using idx_table5_2 on table5 t2  (cost=0.57..18016089.34 rows=1000000064 width=4)  
(6 rows)  
  
postgres=# select count(*) from table5 t1 join table5 t2 using (i) where t1.i<10000000 ;     
  count    
---------  
 9999999  
(1 row)  
  
Time: 3161.559 ms (00:03.162)  
           

https://github.com/digoal/blog/blob/master/201903/20190317_10.md#2%E5%BC%80%E5%90%AF%E5%B9%B6%E8%A1%8C%E8%80%97%E6%97%B6-1-%E7%A7%92 2、开启并行,耗时: 1 秒。

postgres=# explain select count(*) from table5 t1 join table5 t2 using (i) where t1.i<10000000 ;     
                                                            QUERY PLAN                                                              
----------------------------------------------------------------------------------------------------------------------------------  
 Finalize Aggregate  (cost=9785802.62..9785802.63 rows=1 width=8)  
   ->  Gather  (cost=9785802.59..9785802.60 rows=8 width=8)  
         Workers Planned: 8  
         ->  Partial Aggregate  (cost=9785802.59..9785802.60 rows=1 width=8)  
               ->  Merge Join  (cost=1.15..9783168.01 rows=1053830 width=0)  
                     Merge Cond: (t2.i = t1.i)  
                     ->  Parallel Index Only Scan using idx_table5_2 on table5 t2  (cost=0.57..9266088.78 rows=125000008 width=4)  
                     ->  Index Only Scan using idx_table5_2 on table5 t1  (cost=0.57..172964.32 rows=8430637 width=4)  
                           Index Cond: (i < 10000000)  
(9 rows)  
    
postgres=# select count(*) from table5 t1 join table5 t2 using (i) where t1.i<10000000 ;     
  count    
---------  
 9999999  
(1 row)  
  
Time: 1008.615 ms (00:01.009)  
           

https://github.com/digoal/blog/blob/master/201903/20190317_10.md#%E5%85%B6%E4%BB%96%E7%9F%A5%E8%AF%86 其他知识

2、function, op 识别是否支持parallel

postgres=# select proparallel,proname from pg_proc;                                      
 proparallel |                   proname                                                          
-------------+----------------------------------------------                                      
 s           | boolin                                      
 s           | boolout                                      
 s           | byteain                                      
 s           | byteaout                                      
           

3、subquery mapreduce unlogged table

对于一些情况,如果期望简化优化器对非常非常复杂的SQL并行优化的负担,可以自己将SQL拆成几段,中间结果使用unlogged table保存,类似mapreduce的思想。unlogged table同样支持parallel 计算。

4、vacuum,垃圾回收并行。

5、dblink 异步调用并行

《PostgreSQL VOPS 向量计算 + DBLINK异步并行 - 单实例 10亿 聚合计算跑进2秒》 《PostgreSQL 相似搜索分布式架构设计与实践 - dblink异步调用与多机并行(远程 游标+记录 UDF实例)》 《PostgreSQL dblink异步调用实现 并行hash分片JOIN - 含数据交、并、差 提速案例 - 含dblink VS pg 11 parallel hash join VS pg 11 智能分区JOIN》

暂时不允许并行的场景(将来PG会继续扩大支持范围):

1、修改行,锁行,除了create table as , select into, create mview这几个可以使用并行。

2、query 会被中断时,例如cursor , loop in PL/SQL ,因为涉及到中间处理,所以不建议开启并行。

3、paralle unsafe udf ,这种UDF不会并行

4、嵌套并行(udf (内部query并行)),外部调用这个UDF的SQL不会并行。(主要是防止large parallel workers )

5、SSI 隔离级别

https://github.com/digoal/blog/blob/master/201903/20190317_10.md#%E5%8F%82%E8%80%83 参考

https://www.postgresql.org/docs/11/parallel-plans.html 《PostgreSQL 11 preview - 并行计算 增强 汇总》 《PostgreSQL 10 自定义并行计算聚合函数的原理与实践 - (含array_agg合并多个数组为单个一元数组的例子)》

https://github.com/digoal/blog/blob/master/201903/20190317_10.md#%E5%85%8D%E8%B4%B9%E9%A2%86%E5%8F%96%E9%98%BF%E9%87%8C%E4%BA%91rds-postgresql%E5%AE%9E%E4%BE%8Becs%E8%99%9A%E6%8B%9F%E6%9C%BA 免费领取阿里云RDS PostgreSQL实例、ECS虚拟机

PostgreSQL 并行计算解说 之18 - parallel merge join