Veritas InfoScale™ 8.0.2 Storage and Availability Management for Oracle Databases - AIX, Linux, Solaris

To create a clone of an Oracle database by using FlashSnap

1. Create a configuration file.

   $ /opt/VRTS/bin/vxsfadm -s flashsnap \
   -o setdefaults -a oracle --oracle_sid sfaedb \
   --oracle_home /ora_base/db_home -c TEST1
   Written config file TEST1

   This command creates a default configuration file at the location mentioned by the argument given to the -c option. If a path is not specified, then the configuration file is created in the present working directory with all the parameters and default values. You can change the parameters, if required.

   Note:

   If a configuration file with the specified name already exists, then specify a unique configuration name, different path, or remove the existing configuration file. Also, the delete operation does not delete this configuration file.

2. Validate the setup.

   $ /opt/VRTS/bin/vxsfadm -s flashsnap \
   -o validate -a oracle --oracle_sid sfaedb \
   --oracle_home /ora_base/db_home --flashsnap_name \
   TEST1 --app_mode online
   Validating database configuration for third-mirror-break-off snapshot:
   ORACLE_SID = sfaedb
   ORACLE_HOME = /ora_base/db_home
   APP_MODE = online
   SNAPSHOT_ARCHIVE_LOG = yes
   REVERSE_RESYNC_ARCHIVE_LOG = no
   REVERSE_RESYNC_RECOVERY = auto
   ARCHIVELOG_DEST = /arch/sfaedb
   Database validation successful.
   Validating database volume layout for third-mirror-break-off snapshot:
   Data volumes ready for snapshot:
   Volume/volume-set datavol of diskgroup datadg mounted on /data.
   Archivelog volume ready for snapshot:
   Volume/volume-set archvol of diskgroup datadg mounted on /arch.
   Storage units to be used for snapshot from diskgroup datadg:
   vmdk0_25 vmdk0_27
   SNAPSHOT_VOL_PREFIX = SNAP_
   SNAPSHOT_DG_PREFIX = SNAP_
   Database volume layout validated successfully.

   This command validates the configuration file and the database environment. In case of any problem, appropriate error messages are displayed that you can use to correct the problem and then retry.

3. Create a snapshot of the database.

   $ /opt/VRTS/bin/vxsfadm -s flashsnap \
   -o snap  -a oracle --oracle_sid sfaedb --oracle_home \
   /ora_base/db_home --flashsnap_name TEST1 \
   --app_mode online
   Validating database configuration for third-mirror-break-off snapshot:
   ORACLE_SID = sfaedb
   ORACLE_HOME = /ora_base/db_home
   APP_MODE = online
   ARCHIVELOG_DEST = /arch/sfaedb
   SNAPSHOT_ARCHIVE_LOG = yes
   REVERSE_RESYNC_ARCHIVE_LOG = no
   REVERSE_RESYNC_RECOVERY = auto
   Database validation successful.
   snapshot started at Mon Aug 26 10:33:35 2013.
   Putting database in backup mode ... Done
   Validating database volume layout for third-mirror-break-off snapshot:
   Data volumes ready for snapshot:
   Volume/volume-set datavol of diskgroup datadg mounted on /data.
   Archivelog volume ready for snapshot:
   Volume/volume-set archvol of diskgroup datadg mounted on /arch.
   Storage units to be used for snapshot from diskgroup datadg:
   emc0_7 emc0_5
   SNAPSHOT_VOL_PREFIX = SNAP_
   SNAPSHOT_DG_PREFIX = SNAP_
   Database volume layout validated successfully.
   Creating snapshot volumes for data volumes ... Done
   Taking database out of backup mode ... Done
   Flushing online redo logs ... Done
   Creating snapshot volume for archivelog volume ... Done
   Copying snapshot information to snapshot volume ... Done
   Creating snapshot diskgroups ... Done
   Deporting snapshot diskgroups ... Done
   SNAP_datadg
   snapshot ended at Tue Aug 27 00:33:54 2013.

   This command breaks the user-specified mirror ( parameter SNAPSHOT_PLEX_TAG ) from the primary volumes and creates a new disk group with the name starting with the string defined in the snap_dg_prefix parameter. The default value of this parameter is SNAP_.

   When you create a snapshot, the broken-off mirrors are split off into a different disk group that is referred to as the snapshot disk group. The mirrors are the only components that are present in the snapshot disk group. If additional objects, such as cache objects, must be moved from the original disk group to the snapshot disk group, you can use the EXTRA_OBJECTS parameter to specify these additional objects.

   See FlashSnap configuration parameters.

   Note:

   At the validation stage, all the parameters, including the mandatory parameters --oracle_sid and --flashsnap_name, are read and stored in the repository.

   If you need to change any parameter, change the configuration file and specify it with the -c option.

4. Mount the snapshot.

   $ /opt/VRTS/bin/vxsfadm -s flashsnap \
   -o mount -a oracle --oracle_sid sfaedb \
   --oracle_home /ora_base/db_home --flashsnap_name TEST1
   Retrieving snapshot information ...                          Done
   Importing snapshot diskgroups ...                            Done
   Mounting snapshot volumes ...                                Done

   Note:

   This command mounts the snapshot on the host to which storage is attached. The secondary host is the system defined in the SECONDARY_HOST parameter of the configuration file.

   By default, volumes are mounted under the /var/tmp file system.

   If you need to specify an alternate location for mounting snapshot volumes, either provide CLONE_PATH on the command line or from the configuration file.

   For performing off-host operations, specify the SFDB repository host using the -r option of the vxsfadm command.

   $ vxsfadm -s flashsnap -a oracle -o mount \
   --oracle_sid sfaedb --flashsnap_name TEST1 --clone_path /cloneoracle

   Note:

   Ensure that the Oracle user has the required permissions to create the /cloneoracle directory, if it does not exist.

5. Clone the database based on the snapshot.

   $ /opt/VRTS/bin/vxsfadm -s flashsnap -o clone \
   -a oracle --oracle_sid sfaedb --oracle_home /ora_base/db_home \
   --flashsnap_name TEST1 --clone_name clone1
   Retrieving snapshot information ... Done
   Importing snapshot diskgroups ... Done
   Mounting snapshot volumes ... Done
   Preparing parameter file for clone database ... Done
   Mounting clone database clone1 ... Done
   Starting database recovery ...
   Clone database clone1 is open.

   If you have not specified clone_name, it is automatically generated.

   Note:

   If you have already specified the clone_name and the clone_path parameters in the configuration file that was used during the validate operation, the clone_name parameter is not required on the command line.

   If you need to change a memory tuning parameter of the database clone such as memory_target , you can specify a modified pfile during the clone operation.