Home » Server Options » RAC & Failsafe » Parallel query and RAC
Parallel query and RAC [message #74882] |
Tue, 16 March 2004 23:36  |
Mike Wielonski
Messages: 2 Registered: March 2004
|
Junior Member |
|
|
I am seeing some strange behavior with regards to parallel processing in a 2 node cluster and wanted to know if anyone else had seen something similar.� When running a query from node1:
sql> select /*+ parallel(t,4) full(t) */ count(*) from t;
Four�parallel slaves on node2 become active for the duration of the query.� At no time during the above query (via monitoring v$px_process) do any of the parallel slaves on node1 go active.� If the same query is run from node2, it still uses only node2 parallel slaves.� We are running 9.2.0.4 RAC.� The following of both nodes is true:
- this is a test system, therefore my�session is the only one on either cluster
- no parallel slaves are active prior to above sql
- table t has degre=1 and instances=1
- parallel_max_servers=16
- parallel_min_servers=8
- explain plan for query shows intent to use parallel processing
- cpu_count=64
- parallel_automatic_tuning=FALSE
- parallel_adaptive_multiuser=FALSE
- parallel_server=TRUE
Any explanations as to why this might be happening or similar findings would be greatly appreciated.
Thanks,
Mike
|
|
|
|
|
|
|
Goto Forum:
Current Time: Fri May 02 16:38:39 CDT 2025
|