[ monetdb-Bugs-2890035 ] SQL: optimizer confused by concurrency

[SourceForge.net]

Bugs item #2890035, was opened at 2009-10-31 19:21
Message generated for change (Comment added) made by mlkersten
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=482468&aid=2890035&group_id=56967

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: SQL/Core
Group: SQL "stable"
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Wouter Alink (vzzzbx)
Assigned to: Martin Kersten (mlkersten)
Summary: SQL: optimizer confused by concurrency

Initial Comment:
The following happens: (I think on a two weeks old cvs version of Aug2009 branch)

Firing queries at MonetDB using multiple threads (10) in a java program, connecting to MonetDB/SQL via JDBC. Each thread executes a series of queries (50 queries). All threads run the same SQL query (only different literal values). Each thread uses its own JDBC-connection. Each query runs for +-5 to 10 seconds. After a while, 1 thread throws an exception (all others continue):

SQLException:optimizer: 'multiplex' should be the last
SQLException:optimizer: 'deadcode' should be used at least once
SQLException:optimizer: pipeline set to 'inline,remap,evaluate,costModel,coercions,emptySet,aliases,mergetable,deadcode,constants,commonTerms,joinPath,deadcode,reduce,garbageCollector,dataflow,history,multiplex'

This seems to happens consistently when firing queries with 10 threads or more. It does not happen when running with only 4 threads.
What could be wrong? What information is relevant to find the cause?

the queries look like:
START TRANSACTION;
CREATE VIEW ...
CREATE TABLE ...
CREATE VIEW ...
SELECT ...
ROLLBACK;
the database is created from a SQL dump of +-40GB.
the system is loki (a quad-core machine)


----------------------------------------------------------------------

>Comment By: Martin Kersten (mlkersten)
Date: 2009-10-31 19:58

Message:
Patch has been provided on the candidate release (Nov2009)

----------------------------------------------------------------------

Comment By: Martin Kersten (mlkersten)
Date: 2009-10-31 19:44

Message:
Most likely a concurrency conflict on the optimizer pipeline structures.
The pipeline is global and current not protected it seems by a lock.

----------------------------------------------------------------------

You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=482468&aid=2890035&group_id=56967

------------------------------------------------------------------------------
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_______________________________________________
Monetdb-bugs mailing list
Monetdb-bugs@...
https://lists.sourceforge.net/lists/listinfo/monetdb-bugs