0

I am using "PostgreSQL 11.4, compiled by Visual C++ build 1914, 64-bit"

I want to run parallel query for testing purpose, below is my pg_settings parameter values: 

"checkpoint_completion_target"    >> "0.5"  
"default_statistics_target"       >> "100"  
"effective_cache_size"            >> "524288" "8kB"
"maintenance_work_mem"            >> "65536" "kB"
"max_connections"                 >> "100"  
"max_parallel_workers"            >> "8"    
"max_parallel_workers_per_gather" >> "2"    
"max_wal_size"                    >> "1024" "MB"
"max_worker_processes"            >> "8"    
"min_wal_size"                    >> "80" "MB"
"random_page_cost"                >> "4"    
"shared_buffers"                  >> "16384" "8kB"
"wal_buffers"                     >> "512" "8kB"
"work_mem"                        >> "4096" "kB"

While I try to explain the query it doesn't show any 'Workers Planned' how to confirm if my DB support parallel query or not?

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

EXPLAIN select /*+ PARALLEL(test_table 2) */ * from test_table where col_6 = 'Submitted'
--------------------------------------------------------------------
"Seq Scan on test_table  (cost=0.00..3858.50 rows=2633 width=928)"
"  Filter: (col_6 = 'Submitted'::text)"

====================================================================

If i enable force_parallel_mode then it shows 'Workers Planned' but always value with 1. what is the wrong with my setting or DB to run parallel query?

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

set force_parallel_mode = on;
--------------------------------------------------------------------
EXPLAIN select /*+ PARALLEL(test_table 2) */ * from test_table where col_6 = 'Submitted'
--------------------------------------------------------------------
"Gather  (cost=1000.00..5121.80 rows=2633 width=928)"
"  Workers Planned: 1"
"  Single Copy: true"
"  ->  Seq Scan on test_table  (cost=0.00..3858.50 rows=2633 width=928)"
"        Filter: (col_6 = 'Submitted'::text)"

====================================================================

Abdullah-Al Roman
  • 11
  • 1
  • A table with only 2633 rows is probably considered too small for parallel query - check `min_parallel_table_scan_size` Also: `/*+ PARALLEL(test_table 2) */ ` does **nothing** in Postgres –  Jul 03 '19 at 13:47
  • min_parallel_table_scan_size >> "8MB" total table row size is : select count(*) from test_table >> "79240" and 2633 rows is for filtered record. SELECT pg_size_pretty( pg_total_relation_size('test_table') ); >> "22MB" – Abdullah-Al Roman Jul 04 '19 at 10:07
  • See [Laurenz's answer](https://stackoverflow.com/a/56882885/330315) about the formula that is applied when choosing how many workers to use. In your case the result of that formula is 1. –  Jul 04 '19 at 10:11

0 Answers0