2 Upgrading Oracle GoldenGate for Big Data
This chapter describes how to upgrade to Oracle GoldenGate for Big Data 19c (19.1.0.0) by downloading the product as described in Downloading Oracle GoldenGate for Big Data, and then choosing the upgrade paths that suits your environment.
After upgrading, you must then convert your Extract processes to Replicat processes.
Note:
There is no supported upgrade path from release 12.1.2 to 12.2.0.x.
There is no supported upgrade path from release 12.2.0.1.0 to 12.2.0.1.1.
Topics:
2.1 Upgrading by Overwriting the Existing Installation
The most straightforward upgrade path is to copy the Oracle GoldenGate for Big Data 19c (19.1.0.0.x) files into the existing 12c (12.3.1.x or 12.2.0.x) installation directory. Overwriting the product files is possible because there is neither structural nor package name changes in 19.1.0.0.x. In most cases the 12.3.1.x and 12.2.0.x handler and formatter configurations are also compatible with the 19.1.0.0 release. However, is some cases configuration updates may be required.
- (Source systems) Back up the current Oracle GoldenGate for Big Data installation directory on the source system, and any working directories that you have installed on a shared drive in a cluster (if applicable).
- (Source system) Stop user activity on objects in the Oracle GoldenGate configuration.
- Create a new installation directory for the 19.1.0.0 installation with no spaces in its name.
- Extract the ZIP file into this new installation directory, which divides the files into several subdirectories.
- Copy the Oracle GoldenGate for Big Data 19c (19.1.0.0.x) files into the existing 12c (12.3.1.x or 12.2.0.x) installation directory.
- Start the Replicat processes and verify that they are
running.
GGSCI> START MANAGER GGSCI> START REPLICAT group_name GGSCI> INFO REPLICAT group_name GGSCI> VIEW REPORT group_name
Certain 12.2/12.3 JAR files are still present after the overwriting process though the new 19.1.0.0 JAR files are used.
2.2 Upgrading by Installing into a New Directory
Use the following steps to install the Oracle GoldenGate for Big Data 19c (19.1.0.0.x) files into a new installation directory.
- (Source systems) Back up the current Oracle GoldenGate for Big Data installation directory on the source systems, and any working directories that you have installed on a shared drive in a cluster (if applicable).
- (Source system) Stop user activity on objects in the Oracle GoldenGate configuration.
- Create a new installation directory for the 19.1.0.0 installation with no spaces in its name.
- Extract the ZIP file into this new installation directory, which divides the files into several subdirectories.
- Start GGSCI to create the remaining subdirectories in the installation
location.
$ ggsci GGSCI> CREATE SUBDIRS
-
Copy all of the
dirprm
files from your existing installation into thedirprm
directory in the new installation location.Note:
All of your configuration files must be in the
dirprm
directory. If you have property files, Velocity templates, or other configuration files in a location other thandirprm
in your old installation, then you must copy them to thedirprm
directory in the new installation. -
Copy all of the
dirdef
files from your existing installation into thedirdef
directory in the new installation location. -
If you have data files stored in the 12.3.1.x or 12.2.0.x installation
dirdat
directory, then copy or move the existing trail files to thedirdat
directory of the new installation. -
If you have additional JAR files or other custom files in your 12.3.1.x or 12.2.0.x installation, then copy them to the new installation directory.
-
Configure the Replicat processes in the new installation directory by starting GGSCI and adding the Replicat and naming the trails.
GGSCI> ADD REPLICAT group_name, EXTTRAIL trail_name, ... GGSCI> ALTER group_name EXTSEQNO seqno EXTRBA rba
Optionally, you could alter the starting position of Replicat processing as needed.
-
Start the Replicat processes and verify that they are running.
GGSCI> START MANAGER GGSCI> START REPLICAT group_name GGSCI> INFO REPLICAT group_name GGSCI> VIEW REPORT group_name
-
Modify the source system to write to the new Oracle GoldenGate for Big Data installation directory:
-
(Optional) Upgrade the source database Oracle GoldenGate capture following the upgrade procedure for your database platform.
-
Configure the source database capture to write to the new Oracle GoldenGate for Big Data 12.3.1.x installation
dirdat
directory. -
When the old Oracle GoldenGate for Big Data installation has processed all its data, switch over to the process that will send data to the new location.
-
2.3 Switching Existing Extract Processes to Replicat Processes
In previous releases, you could use an Extract and pump process to write to your Big Data targets. In this release, this solution is deprecated so you must use Replicat.
A typical Extract configuration is similar to:
EXTRACT mygroup
SOURCEDEFS path/to/source/def/file
CUSEREXIT libggjava_ue.so CUSEREXIT PASSTHRU INCLUDEUPDATEBEFORES
GETUPDATEBEFORES
TABLE *.*;
With Replicat the preceding configuration would be:
REPLICAT mygroup
SOURCEDEFS path/to/source/def/file
TARGETDB LIBFILE libggjava.so SET property=/path/to/properties/file
MAP *.* TARGET *.*;
The same properties file works in the Replicat configuration so you do not need to change the properties file.
To complete this process, you must add the Replicat group, start the process, and verify that it is running:
GGSCI> ADD REPLICAT group_name, EXTTRAIL trail_name, ...
GGSCI> ALTER group_name EXTSEQNO seqno EXTRBA rbaGGSCI> START REPLICAT group_name
GGSCI> INFO REPLICAT group_name
GGSCI> VIEW REPORT group_name