Database Reference
In-Depth Information
BEGIN
DBMS_RESOURCE_MANAGER.CLEAR_PENDING_AREA();
END;
/
This will save all the ODRM definitions created in the workspace area to disk.
In the preceding definition, we have three groups for which plans are defined. Group OTHER_GROUP is an
Oracle-provided default group for every resource plan definition. The
HIGH_PRIORITY
and
LOW_PRIORITY
are groups
created based on Table
15-5
in Step 7.2 previously. Based on the application distribution in Table
15-5
, GRUD is
defined under resource group
LOW_PRIORITY
(running under low priority) on instances
SCDB_1
,
SCDB_2
, and
SCDB_3
.
Application TAPS and TICKS are defined under resource group
HIGH_PRIORITY
and share instances with application
GRUD. Based on the requirements, the resource plan shares the resources between the two resource groups
HIGH_PRIORITY
and
LOW_PRIORITY
, giving resource group
HIGH_PRIORITY
more resources.
The default group,
OTHER_GROUP
, should not be ignored. At times, when there are runaway processes and both
resource groups consume all the resources, it would be in the best interest of the DBA to allocate some resources
under the
OTHER_GROUP
category to allow the DBA to interview and perform any administrative operation.
Step 7.6
The consumer group and priority definitions could be verified by querying against the
DBA_RSRC_GROUP_MAPPINGS
view:
SELECT ATTRIBUTE,
VALUE,
CONSUMER_GROUP
FROM DBA_RSRC_GROUP_MAPPINGS
WHERE ATTRIBUTE LIKE '%SERVICE%';
ATTRIBUTE VALUE CONSUMER_GROUP
------------------------ -------- ----------------
SERVICE_NAME TAPS HIGH_PRIORITY
SERVICE_NAME GRUD LOW_PRIORITY
SERVICE_NAME FIPS STD_PRIORITY
SERVICE_NAME SSKY STD_PRIORITY
Step 7.7 Job Class Definition
One application listed in Table
15-5
is a batch job (reporting) that is triggered by other applications on other
services in the cluster. Batch jobs are normally scheduled to run at predefined intervals and a predefined frequency.
DBMS_SCHEDULER
can schedule a batch job. One prerequisite to define a batch job using the
DBMS_SCHEDULER
is to
define a job class using the
CREATE_JOB_CLASS
procedure:
EXECUTE DBMS_SCHEDULER.CREATE_JOB_CLASS -
(JOB_CLASS_NAME => 'SSKY', -
RESOURCE_CONSUMER_GROUP => NULL, -
SERVICE=> 'SSKY', -
LOGGING_LEVEL=> DBMS_SCHEDULER.LOGGING_RUNS,-
LOG_HISTORY => 30);
Search WWH ::
Custom Search