PostgreSQL 并行計算解說 之12 - parallel in rc,rr 隔離級别

标簽

PostgreSQL , cpu 并行 , smp 并行 , 并行計算 , gpu 并行 , 并行過程支援

https://github.com/digoal/blog/blob/master/201903/20190317_04.md#%E8%83%8C%E6%99%AF 背景

PostgreSQL 11 優化器已經支援了非常多場合的并行。簡單估計，已支援27餘種場景的并行計算。

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 自定義并行聚合                        
                        
parallel 自定義并行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                         
           

接下來進行一一介紹。

關鍵知識請先自行了解：

1、優化器自動并行度算法 CBO

《PostgreSQL 9.6 并行計算 優化器算法淺析》 《PostgreSQL 11 并行計算算法，參數，強制并行度設定》

https://github.com/digoal/blog/blob/master/201903/20190317_04.md#parallel-in-rcrr-%E9%9A%94%E7%A6%BB%E7%BA%A7%E5%88%AB parallel in rc,rr 隔離級别

并行計算支援rc,rr隔離級别，暫時未支援ssi隔離級别。

資料量：10億。

場景	資料量	關閉并行	開啟并行	并行度	開啟并行性能提升倍數
rc (ud agg count distinct)	10 億	107 秒	3.65 秒	30	29.3 倍
rr (ud agg count distinct)

https://github.com/digoal/blog/blob/master/201903/20190317_04.md#1%E5%85%B3%E9%97%AD%E5%B9%B6%E8%A1%8C%E8%80%97%E6%97%B6-107-%E7%A7%92 1、關閉并行，耗時： 107 秒。

postgres=# begin isolation level read committed ;  
BEGIN  
postgres=# explain select count_distinct(i),count(i) from table1;  
                                 QUERY PLAN                                   
----------------------------------------------------------------------------  
 Aggregate  (cost=19424779.00..19424779.01 rows=1 width=16)  
   ->  Seq Scan on table1  (cost=0.00..14424779.00 rows=1000000000 width=2)  
(2 rows)  
  
postgres=# begin isolation level repeatable read ;  
BEGIN  
postgres=# explain select count_distinct(i),count(i) from table1;  
                                 QUERY PLAN                                   
----------------------------------------------------------------------------  
 Aggregate  (cost=19424779.00..19424779.01 rows=1 width=16)  
   ->  Seq Scan on table1  (cost=0.00..14424779.00 rows=1000000000 width=2)  
(2 rows)  
  
  
postgres=# begin isolation level read committed ;  
BEGIN  
postgres=# select count_distinct(i),count(i) from table1;  
 count_distinct |   count      
----------------+------------  
              1 | 1000000000  
(1 row)  
  
Time: 107127.119 ms (01:47.127)  
  
postgres=# begin isolation level repeatable read ;  
BEGIN  
postgres=# select count_distinct(i),count(i) from table1;  
 count_distinct |   count      
----------------+------------  
              1 | 1000000000  
(1 row)  
  
Time: 106633.829 ms (01:46.634)  
           

https://github.com/digoal/blog/blob/master/201903/20190317_04.md#2%E5%BC%80%E5%90%AF%E5%B9%B6%E8%A1%8C%E8%80%97%E6%97%B6-365-%E7%A7%92 2、開啟并行，耗時： 3.65 秒。

postgres=# begin isolation level read committed ;  
BEGIN  
postgres=# explain select count_distinct(i),count(i) from table1;  
                                          QUERY PLAN                                            
----------------------------------------------------------------------------------------------  
 Finalize Aggregate  (cost=4924779.24..4924779.25 rows=1 width=16)  
   ->  Gather  (cost=4924779.00..4924779.01 rows=30 width=40)  
         Workers Planned: 30  
         ->  Partial Aggregate  (cost=4924779.00..4924779.01 rows=1 width=40)  
               ->  Parallel Seq Scan on table1  (cost=0.00..4758112.33 rows=33333333 width=2)  
(5 rows)  
  
postgres=# begin isolation level repeatable read ;  
BEGIN  
  
postgres=# explain select count_distinct(i),count(i) from table1;  
                                          QUERY PLAN                                            
----------------------------------------------------------------------------------------------  
 Finalize Aggregate  (cost=4924779.24..4924779.25 rows=1 width=16)  
   ->  Gather  (cost=4924779.00..4924779.01 rows=30 width=40)  
         Workers Planned: 30  
         ->  Partial Aggregate  (cost=4924779.00..4924779.01 rows=1 width=40)  
               ->  Parallel Seq Scan on table1  (cost=0.00..4758112.33 rows=33333333 width=2)  
(5 rows)  
  
  
postgres=# begin isolation level read committed ;  
BEGIN  
postgres=# select count_distinct(i),count(i) from table1;  
 count_distinct |   count      
----------------+------------  
              1 | 1000000000  
(1 row)  
  
Time: 3654.470 ms (00:03.654)  
  
postgres=# begin isolation level repeatable read ;  
BEGIN  
postgres=# select count_distinct(i),count(i) from table1;  
 count_distinct |   count      
----------------+------------  
              1 | 1000000000  
(1 row)  
  
Time: 3658.730 ms (00:03.659)  
           

https://github.com/digoal/blog/blob/master/201903/20190317_04.md#%E5%85%B6%E4%BB%96%E7%9F%A5%E8%AF%86 其他知識

2、function, op 識别是否支援parallel

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

3、subquery mapreduce unlogged table

對于一些情況，如果期望簡化優化器對非常非常複雜的SQL并行優化的負擔，可以自己将SQL拆成幾段，中間結果使用unlogged table儲存，類似mapreduce的思想。unlogged table同樣支援parallel 計算。

4、vacuum，垃圾回收并行。

5、dblink 異步調用并行

《PostgreSQL VOPS 向量計算 + DBLINK異步并行 - 單執行個體 10億 聚合計算跑進2秒》 《PostgreSQL 相似搜尋分布式架構設計與實踐 - dblink異步調用與多機并行(遠端 遊标+記錄 UDF執行個體)》 《PostgreSQL dblink異步調用實作 并行hash分片JOIN - 含資料交、并、差 提速案例 - 含dblink VS pg 11 parallel hash join VS pg 11 智能分區JOIN》

暫時不允許并行的場景(将來PG會繼續擴大支援範圍)：

1、修改行，鎖行，除了create table as , select into, create mview這幾個可以使用并行。

2、query 會被中斷時，例如cursor , loop in PL/SQL ，因為涉及到中間處理，是以不建議開啟并行。

3、paralle unsafe udf ，這種UDF不會并行

4、嵌套并行(udf (内部query并行))，外部調用這個UDF的SQL不會并行。（主要是防止large parallel workers ）

5、SSI 隔離級别

https://github.com/digoal/blog/blob/master/201903/20190317_04.md#%E5%8F%82%E8%80%83 參考

https://www.postgresql.org/docs/11/parallel-plans.html 《PostgreSQL 11 preview - 并行計算 增強 彙總》 《PostgreSQL 10 自定義并行計算聚合函數的原理與實踐 - (含array_agg合并多個數組為單個一進制數組的例子)》

https://github.com/digoal/blog/blob/master/201903/20190317_04.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 免費領取阿裡雲RDS PostgreSQL執行個體、ECS虛拟機