HP 3PAR Application Software Suite for Oracle - Manual (Page 8)

where:
•
<db_server> is the host name of the database server.
•
<oracle_sid> is the Oracle SID of the database instance.
The following example displays the location of the Recovery Manager for Oracle repository on
the backup server for Oracle database instance test that is running on database server Host1.
/etc/3par/solutions/Host1.ora.test
If the database is a Real Application Cluster (RAC) database, there will be multiple repositories,
one for each RAC instance.
NOTE: For more information about using the utility to manage the repository, refer to “The Virtual
Copy Repository” (page 16) .
Interacting with Oracle
HP 3PAR Recovery Manager for Oracle interacts with Oracle database through the SQL*Plus utility
to perform the following:
•
Retrieve database structure information in order to create Virtual Copy for the database.
•
Interact with the Oracle database (putting database in backup mode, stopping redo applied
process or performing database log switching), as necessary, to create a consistent Virtual
Copy.
NOTE: Recovery Manager for Oracle supports only Oracle Enterprise Edition. It does not
support Oracle Standard Edition.
To create a consistent Virtual Copy of an Oracle database, the database structure must satisfy the
following requirements:
•
The database must be running in archive log mode and automatic archiving must be enabled
to create an online Virtual Copy, datafile Virtual Copy, or archive log Virtual Copy.
•
Datafiles and archive logs must reside on separate HP 3PAR virtual volumes.
•
The online redo logs and control files should not reside on the virtual volumes used by the
datafiles and archive logs to avoid being rolled back along with datafiles and archive logs
virtual volumes. However, the online redo logs and control files can share the same HP 3PAR
virtual volumes.
•
If the database files reside on Symantec VxVM volumes, datafiles and archive logs must reside
on separate VxVM disk groups. The online redo logs and control files should reside on separate
VxVM volumes used by the datafiles and archive logs.
•
If the Oracle database is an ASM managed database, the datafiles and archive logs must
reside on separate ASM disk groups. The online redo logs and control files should not reside
on the same ASM disk groups used by the datafiles and archive logs to avoid being rolled
back when using the Recovery Manager Rollback feature.
•
ASM disk groups should not be shared between different databases.
•
If using Logical Volume Manager (LVM) on HP, Linux, or AIX, the Oracle datafiles and archive
logs must reside on separate LVM volume groups. In addition, online redo logs and control
files must not reside on LVM volume groups that are used by Oracle datafiles and archive
logs. However, the online redo logs and control files can reside on the same LVM volume
group.
•
If the database is an RAC database, all RAC instances must share the same archive log
destinations (same cluster file systems or same ASM disk groups).
8 Overview of Recovery Manager Operations