RE: PARALLEL_EXECUTION_MESSAGE_SIZE
- From: "Pass, Stephanie" <Stephanie_Pass@xxxxxxxxxxxxxxxxx>
- To: <Steve.Rospo@xxxxxxxxxxx>, "Oracle-L" <oracle-l@xxxxxxxxxxxxx>
- Date: Thu, 6 Oct 2005 08:48:33 -0600
Years ago when I was helping a group move from an OPS database to a Sequent, we
ran into issues with PQO. These are some of my notes I took when watching the
system. Maybe they will help you. Take it with a grain of salt the db version
was - 8.1.5.
Info on each PX message wait:
Currently waiting on:PX Deq: Execution Msg -
PQ slave is waiting to be told what to do.
This is one of the main events used in a parallel query
dialogue as the slave is told to parse / execute / fetch etc..
Currently waiting on:PX Deq: Join ACK -
When setting up PQ groups of slaves we wait for them to
join the group.
Currently waiting on:PX Deq: Execute Reply -
QC is waiting for a reply
Currently waiting on:PX Deq: Signal ACK -
Dependency: We are dependent on the producer.
Advise: If the process is often waiting for data then look at the
execution plan/s and degree of parallelism.
Currently waiting on:PX qref latch -
Each process queue has an associated qref latch, which needs
to be acquired before the queue buffers can be manipulated.
Currently waiting on:PX Deq: Table Q Normal
Currently waiting on:latch free
It seems that we may be falling into this problem:
PARALLEL QUERY DOES NOT SCALE IN 8.1 WHEN # OF DISKS IS SMALL RELATIVE TO
DEGREE
-----Original Message-----
From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx]On
Behalf Of Steve Rospo
Sent: Wednesday, October 05, 2005 2:51 PM
To: Oracle-L
Subject: RE: PARALLEL_EXECUTION_MESSAGE_SIZE
I found this page which was very interesting:
http://doug.burns.tripod.com/px8.html
"Partial Message Event
* PX Deq: Msg Fragment
--
//www.freelists.org/webpage/oracle-l
Other related posts:
