Have you tried running SQL trace (10046)? Oracle support search yields
the following information:
Bug 12842402 "enq: SV - contention" using ORDERED Sequences in RAC
So, are your sequences ordered? Ordered sequences have been a bane of
RAC application since the time immemorial, since Oracle 8i and OPS.
Also, do you have any enq HW waits? It's a high water mark wait, a sign
that your sequence cache is too small. In any case, I would run 10046
trace on the application. AWR and oracle-l are not sufficient in this case.
Regards
On 6/14/2016 5:14 AM, Pawel Smolarz wrote:
Hi,
Hi,
On Exadata RAC database 12.1.0.2.0 we have problems with queries:
SELECT schema.sequence_name.NEXTVAL
FROM DUAL CONNECT BY ROWNUM <= :v1
huge waits on Foreground Events:
Event Waits Total Wait Time (sec) Wait Avg(ms) % DB time Wait Class
DB CPU 2.1M 37.5
resmgr:cpu quantum 1.1E+08 1.2M 11.21 21.6 Scheduler
enq: SV - contention 83,224,234 1.1M 12.91 18.8 Other
I know that is BUG but fix has been included in 12.1.0.1
Anyone has a experience with this type of event?
A lot of CPU has been consume by these sessions .
Regards,
Pawel
--
//www.freelists.org/webpage/oracle-l
