Resource Manager in highly transactional database?
- From: "Thomas Jeff" <jeff.thomas@xxxxxxxxxxx>
- To: <oracle-l@xxxxxxxxxxxxx>
- Date: Tue, 16 Nov 2004 13:17:42 -0500
We have a 9.2.0.4/AIX 5.2 shipping database that experiences periodic
freezes,
during "waves" when ship/print runs are generated. During these
freezes,
we typically see between 100-200 sessions waiting on various resmgr
waits
(fragment below), for up 4-5 minutes before they eventually clear out.
Is there any tuning/tweaking I can do (other then disabling RM) to get
around this? I did not define the plans, and the intent is that they
are set up more
to faciliate chargeback (via periodic scans of v$rsrc_consumer_group)
then
intended for actual resource control. Checking Metalink/Google has not
come
up with anything useful.
SID Username EVENT STATUS
SECONDS_IN_WAIT
----- ------------ ---------------------------- ----------
---------------
590 OPS$TCESMISP resmgr:wait in actses run WAITING 7
591 OPS$TCESMISP resmgr:wait in actses run WAITING 4
592 OPS$TCESMISP resmgr:wait in actses run WAITING 7
593 OPS$TCESMISP resmgr:waiting in end wait WAITING 26
594 OPS$U54627 resmgr:waiting in end wait WAITING 81
596 OPS$TCESMISP resmgr:waiting in end wait WAITING 29
598 OPS$TCESMISP resmgr:wait in actses run WAITING 3
599 OPS$U21801 resmgr:waiting in end wait WAITING 26
600 OPS$TCESMISP resmgr:wait in actses run WAITING 1
601 OPS$TCESMISP resmgr:wait in actses run WAITING 2
602 OPS$U60967 resmgr:wait in actses run WAITING 47
--------------------------------------------
Jeffery D Thomas
DBA
Thomson Information Services
Thomson, Inc.
Email: jeff.thomas@xxxxxxxxxxx
Indy DBA Master Documentation available at:
http://gkmqp.tce.com/tis_dba
--------------------------------------------
--
//www.freelists.org/webpage/oracle-l
Other related posts:
- » Resource Manager in highly transactional database?
