parallel_degree_policy tips

Question:  Oracle 11g release 2 has introduced a new parallel parameter called parallel_degree_policy.  Is parallel_degree_policy like the parallel_automatic_tuning parameter?

Answer:  The parallel_degree_policy parameter is the successor to the deprecated parallel_automatic_tuning parameter, a parameter with significant issues. The parallel_degree_policy is indeed a replacement for PAT that is designed for use exclusively in data warehouse and decision support environments. (Automatic parallelism is not recommended for OLTP databases or systems with first_rows_n optimization).

Also see the Oracle 12c parallel_degree_level.

• New "augmented" Buffering:  With parallel_degree_policy=auto Oracle can use their new augmented data buffer method dubbed "in-memory parallel execution" for partitioned table scans.  This is a replacement to the PGA-based buffering that was used in parallel large-table full-table scans in non-RAC systems to keep the standard SGA data buffer pool from being overwhelmed since only objects that consume less than 2% of the db_cache_size or db_nnk_cache_size would use the regular SGA buffer.
  
  
• Automatic parallelism and degree:  Setting parallel_degree_policy=auto allows Oracle to decide how much parallelism (if any) to use on a SQL query.  Most shops will hardcode their degree of parallelism based on their own experience with job throughput, a more accurate approach than using any automated solution.
  
  
• SQL queuing:  Setting parallel_degree_policy=auto also enabled parallel query queuing, a feature whereby a SQL is delayed when running it would saturate the I/O subsystem.  This is useful for nighttime batch processing, but its not appropriate for any jobs where minimizing response time is a primary goal.  Most shops use dbms_scheduler to single-thread or parallelize their job manually, in most cases a better approach than an automated solution.

The parallel_degree_policy parameter is related to the amount of table data residing in the data buffer cache. Using parallel_degree_policy allows Oracle to bypass direct path reads when Oracle determines that lots of the table data blocks already reside in the data buffer cache.

In traditional 32-bit systems (limited by on ly a few gig of RAM for the SGA), direct path reads (which bypass the SGA were always faster than reading a large table through the data buffer. However, with the advent of 64-bit servers with dozens of gigabytes for the db_cache_size, large tables are often be fully cached, negating the need to always perform direct path reads for parallel large-table full-table scans.

Guy Harrison conducted some benchmark tests of parallel_degree_policy and we see details on how parallel_degree_policy evaluates the caching of large tables:

"If PARALLEL_DEGREE_POLICY is set to AUTO then Oracle might perform buffered IO instead of direct path IO. . .

The documentation says that the optimizer decides whether or not to use direct path depending on the size of the table and the buffer cache and the likelihood that some data might be in memory."

 

Get the Complete Oracle SQL Tuning Information  The landmark book "Advanced Oracle SQL Tuning  The Definitive Reference"  is filled with valuable information on Oracle SQL Tuning. This book includes scripts and tools to hypercharge Oracle 11g performance and you can buy it for 30% off directly from the publisher.