AWR Top timed events : PX Nsq: PQ load info query & Streams AQ: qmn coordin
subhavsaMar 19 2013 — edited Mar 28 2013Hi,
We are reviewing an AWR report for couple of days. And the below events are consistently reported as Top timed events.
DB CPU time.
enq: TM – contention
Streams AQ: qmn coordinator waiting for slave to start
enq: TX - row lock contention
PX Nsq: PQ load info query
The environment is 2 node RAC on 11.2.0.3.
- For "DB CPU time", "enq: TX - row lock contention" and "enq: TM – contention", we have found couple of recommendations and approach for review.
I have got a question of the other 2 events.
1. "Streams AQ: qmn coordinator waiting for slave to start"
====================================================
Based on metalink and couple of forums, this seems to be linked to aq_tm_processes.
The documents suggests to unset this parameter for 10g onwards. And for 11g, the default value (when it is unset) shows the value of 1 in v$parameter. And that is what is set for our environment and that would mean oracle is managing it automatically. But we still see this event all the time.
Can the aq_tm_processes parameter value be set for 11gR2? If yes, what would be the appropriate value?
( Note: Some documents (Note:746313.1) suggests, in 9.2 and 10g+ databases, AQ_TM_PROCESSES must not be set to 10, as setting it to 10 disables some other features of the Queue Monitoring )
2. PX Nsq: PQ load info query
==========================
Is there anyway to find the root cause of the above wait event? Cannot find much information about the above event.
Would appreciate any recommendation to debug this event
Thanks.
Sunil
Edited by: subhavsa on Mar 19, 2013 11:52 AM