Load not evenly distributed among PX slaves in 11.2
(Doc ID 1224903.1)
Last updated on JUNE 05, 2019
Applies to:Oracle Database - Enterprise Edition - Version 18.104.22.168 and later
Oracle Database Cloud Schema Service - Version N/A and later
Oracle Database Exadata Express Cloud Service - Version N/A and later
Oracle Database Exadata Cloud Machine - Version N/A and later
Oracle Cloud Infrastructure - Database Service - Version N/A and later
Information in this document applies to any platform.
NOTE: In the images and/or the document content below, the user information and data used represents fictitious data from the Oracle sample schema(s) or Public Documentation delivered with an Oracle database product. Any similarity to actual persons, living or dead, is purely coincidental and not intended in any manner.
Load distribution among slaves is fine in 10.2, gets a bit worse in 11.1 and is quite bad in 11.2, causing one slave gets most of the load while the other slaves get just few rows.
That causes performance issues, as one slave takes longer to finish (as per bigger load) and causes overall performance on the query is bigger. As we can check from v$pq_tqstat output.
* 10.2 load balance is quite good
The primary change between 11.2 and 11.1 is that in 11.2 we do granule batching. That is when the slaves request granules, the QC provides them a set of granules instead of a single granule to work on.
The parameter controlling this feature is _px_granule_batch_size with default value of 10 i.e. 10 granules are given to slaves on one request.
To view full details, sign in with your My Oracle Support account.
Don't have a My Oracle Support account? Click to get started!
In this Document