Re: [Postgres-xc-bugs] Coordinator never ends when executing multiple Join queries

[ZhangJulian <jul...@ou...>]

When the query was running 3 hours, had you monitored the CPU and IO stats? Was it locked or running? 

Could you run 
EXPLAIN VERBOSE select.....
to see what the remote query looks like? 

Thanks
Julian
Date: Fri, 11 Jul 2014 16:58:41 -0430
From: fab...@gm...
To: pos...@li...
Subject: [Postgres-xc-bugs] Coordinator never ends when executing multiple	Join queries

Queries with multiple Joins are giving me trouble. My postgres-xc configuration is as follows:
2 datanodes,1 coordinator,1 gtm
Each on a different server. I have 5 tables loaded equally on both nodes: 1 replicated and 4 distributed by hash on the correspondent Join attributes.

When I run the query on a simple PostgreSQL server, it takes about 20 seconds to finish the execution, but it seems like it will never come to an end whenever I run the query on my postgres-xc server; I have waited over three hours waiting for it to end. Nevertheless, if I do "EXECUTE DIRECT ON node# 'query'", it takes up to 5 seconds tops on both nodes.

Suppose my tables are...tableR                            --- Replicated table  
table1                            --- distributed by hash on (id)table2, table3 & table4  --- distributed by hash on (t1id)

... And:table1 has 2200000 recordstable2 has 2160000 recordstable3 has 2190000 recordstable4 has 6596 recordstableR has 223 records


The query is basically something like this:
		
	
	
		
			
				
					SELECT COUNT(*) FROM table1JOIN table2 ON table2.t1id = table1.id

LEFT OUTER JOIN table3 ON table3.t1id = table1.id LEFT OUTER JOIN table4 ON table4.t1id = table1.id
LEFT OUTER JOIN tableR ON  tableR.id = table4.tRid;

Thank you very much,Fabiola Rosato.

				
			
		

------------------------------------------------------------------------------
_______________________________________________
Postgres-xc-bugs mailing list
Pos...@li...
https://lists.sourceforge.net/lists/listinfo/postgres-xc-bugs