Reconfigure an Installation

6 Reconfigure an Installation

After the initial Oracle Big Data SQL installation, there may be changes on the Hadoop cluster or the Oracle Database system that require corresponding updates to the Oracle Big Data SQL configuration, including:

A change in the status of Kerberos — if it was previously disabled and is now enabled, or vice versa.
Implementation of TLS (Transport Layer Security) or other security measures, such as Database Authentication option now supported by Oracle Big Data SQL.
Changes to the location of Hive service instances within the Hadoop cluster.
A switch from Ethernet to InfiniBand (or InfiniBand to Ethernet) for networking between the Hadoop cluster and the Oracle Database system.
A change to the version of Oracle Big Data SQL running on the Hadoop cluster.

Important:

If you are storing Oracle tablespaces in HDFS, then after you run reconfigure , take the tablespaces offline (or stop the database) and then restart the FUSE-DFS service running on all database compute nodes that are connected to the Hadoop cluster. See Restarting FUSE-DFS After a Reconfigure Operation at the end of this section for details.

The basic procedure for reconfiguring an installation is:

Edit the bds-config.json file as needed.
Run the Jaguar reconfigure command to push the configuration changes to the Hadoop cluster nodes and to generate an updated database-side installation bundle.
Reinstall Oracle Big Data Data SQL on Oracle Database compute nodes using the new installation bundle. You can just overwrite the current installation. There is no prerequisite to uninstall the current Oracle Big Data SQL installation on either side.

Details are provided below.

See Also:

The addition or removal of Oracle Big Data SQL from Hadoop cluster nodes or nodes in a multinode Oracle Database system is not a considered configuration change. See Expanding or Shrinking an Installation.

Run “jaguar reconfigure” to Reconfigure the Hadoop Side of the Installation and to Generate an Updated Database Installation Bundle

Log on to the cluster management server node as root and cd to the directory where you extracted the downloaded Oracle Big Data SQL installation bundle.
Go to the BDSJaguar-4.1.1 subdirectory under the path where you unzipped the Oracle Big Data SQL bundle .
```
# cd <Big Data SQL Install Directory>/BDSJaguar-4.1.1
```
In the BDSJaguar directory, run the Jaguar utility. Pass it the reconfigure parameter and the configuration file name (bds-config.json or other) as arguments. You may or may not need to include the --requestdb option.
```
[root@myclusteradminserver:BDSjJaguar] #  ./jaguar reconfigure bds-config.json
```
If bds-config.json sets database_auth_enabled to “true”, then include the --requestdb option in the install command:
```
[root@myclusteradminserver:BDSJaguar] # ./jaguar --requestdb testdb reconfigure
```
The script prompts for the cluster management service administrator credentials and then reconfigures Oracle Big Data SQL on the management server and the cluster nodes according to the new settings in the configuration file. The script output terminates with the following message if the installation completed without error.
```
BigDataSQL: RECONFIGURE workflow completed. 
```

Copy the Updated Database Installation Bundle to Each Oracle Database Node and Install It

Log on as either root or oracle on the cluster management server and cd to the directory where the new database bundle was generated.

# cd <Big Data SQL Install Directory>/BDSJaguar/db-bundles

Check that the new bundle is there.

[oracle@myclusteradminserver:db-bundles] # ls 
[oracle@myclusteradminserver:db-bundles] # bds-4.1.1-db-<cluster>-<yymmdd.hhmi>.zip

Copy the database installation bundle over to any directory on the Oracle Database node, using the oracle account on the Oracle system. If there are multiple bundles, be sure to select the correct bundle for cluster that you want to connect to Oracle Database.
```
[oracle@myclusteradminserver:db-bundles] # scp bds-4.1.1-db-<cluster>-<yymmdd.hhmi>.zip oracle@<database_node>:/opt/tmp
```
Log on to the Oracle Database node as the database owner (oracle or other) in order to perform the installation.
Cd to the directory where you copied the bundle, unzip the bundle, and then run the compressed executable that was extracted from the zip file. For example:
```
$ unzip bds-4.1.1-db-cdh510-170309.1918.zip
$ ./bds-4.1.1-db-cdh510-170309.1918.run
```
Change directories to the new directory under ORACLE_HOME and run bds-database-install.sh, the database-side installer.
```
$ cd $ORACLE_HOME/BDSJaguar/cdh510-6-node1.my.<domain>.com
$ ./bds-database-install.sh
```

Restarting FUSE_DFS After a Reconfigure Operation

The Oracle Tablespaces in HDFS feature uses a FUSE-DFS connection for file transfers between Oracle Database and the Hadoop cluster. There is a FUSE-DFS service instance for each connection to cluster. The service is named as follows:

 fuse_<cluster name>_hdfs

If this service exists, then after a reconfiguration of an Oracle Big Data SQL installation, do the following:

Either set the tablespaces in HDFS to offline or stop the database.

The datafiles for these tablespaces are open and locked by database processes.
Restart FUSE-DFS. Do this on each compute node.
```
# service fuse_mycluster_hdfs restart
```
After FUSE-DFS has restarted, you can set the tablespaces to online (or you if you stopped the database, you can restart it).

The status, stop, and start operations are also available for the FUSE-DFS service, as in:

# service fuse_mycluster_hdfs status

6.1 Reconfiguring an Existing Oracle Big Data SQL Installation

When Oracle Big Data SQL is installed on the cluster management server, the database-side installation bundle that is generated contains configuration information that is based on the current state of the cluster. When you install the bundle on the Oracle Database side, this configuration information is used to set up the connnection between the Hadoop cluster and the database server. After the installation, changes in the Hadoop cluster may require an update to the Oracle Big Data SQL configuration both sides (Hadoop cluster and database server) in order to maintain connectivity.

Here are some Hadoop-side changes which require you to reconfigure the installation:

Network security changes, such as a switch from HTTP to HTTPS or vice versa.
Port reassignments in the cluster management service (CM or Ambari).
Hive changes, such as migration of the service to different node or nodes or a change to the authentication method.
Changes to other related Hadoop services, such as HDFS, MapReduce2, and YARN, may also affect connectivity.

Note:

The reconfigure procedure cannot be used to update the installation of an older version of Oracle Big Data SQL. You should instead overwrite the old installation with the current Oracle Big Data SQL version.

As root on the Hadoop cluster management server, make any necessary updates to the configuration file for the connection between the Hadoop cluster and Oracle Database (bds-config.json or other).
Go to the BDSJaguar directory and run Jaguar to invoke the reconfigure operation. Pass in the configuration filename.
```
# cd <Big Data SQL install directory>/BDSJaguar
# ./jaguar reconfigure bds-config.json
```
This updates the Oracle Big Data SQL installation on the Hadoop cluster and generates a new database-side installation bundle, which you can find at <Oracle Big Data SQL Install Directory>/BDSJaguar/db-bundles.

Note:
If you are using Database Authentication and already provided a request key to the database-side installer when you did the original installation, then you do not need to do again. That is, you do not need to include --requestdb in a reconfigure operation unless you want to set up a connection to another database.
As in the original installation, copy the new bundle file to a temporary directory each database node. Connect using the oracle account on the database node. For example:
```
# scp bds-4.1.1-db-<cluster>-<yymmdd.hhmi>.zip oracle@dbnode:/opt/tmp
```
Log on to the database node as the oracle user.
Locate the bundle and unzip it.
Run the executable file extracted from the bundle.
```
[oracle@mydbserver: $ bds-4.1.1-db-<cluster>-<yymmdd.hhmi>.run
```
This extracts the installation files into a new directory under ORACLE_HOME.
Change directories to ORACLE_HOME and list the sub directories. You should see the old installation directory along with the new one. The directory names will include the same cluster name, but will have different timestamps. Delete the old directory.
```
$ cd $ORACLE_HOME 
$ ls
  bds-4.1.1-db-<cluster>-<new timestamp>
  bds-4.1.1-db-<cluster>-<old timestamp>
$ rm -rf bds-4.1.1-db-<cluster>-<old timestamp>
```
Change directories to the new directory.

Run bds-database-install.sh with the --reconfigure parameter.
```
$ ./bds-database-install.sh --reconfigure
```