Help us improve your experience.

Let us know what you think.

Do you have time for a two-minute survey?

 

Upgrading the Data (D) Nodes with the New Software

 

After all M, SM, and S nodes in the cluster are upgraded, you need to upgrade the software in D nodes by installing the complete SBR release distribution package or by installing only the latest MySQL package.

Installing the Complete SBR Release Distribution Package

Installing the Complete SBR Release Distribution Package

This section describes the procedure to install the new software on the data nodes by copying the complete SBR release distribution package.

To install the new software on the data node by copying the complete SBR release distribution package:

Note

You must repeat this procedure on every D node in the cluster, one at a time. This example procedure uses two D nodes, bng-sbr-perfm3000-3 (10.212.10.188) and bng-sbr-perf2 (10.212.10.67).

  1. Verify that the /opt/JNPRmysql directory contains the directories install and mysql-cluster-advanced-7.4.15-solaris10-sparc-64bit. If these directories are not present, your installation is using non-default locations for these files. Locate these files and execute the steps in the following procedure in the same directory that contains these files.

  2. Log in to the first D node as root, and stop the SSR process. In the following example, the D node is called bng-sbr-perfm3000-3 (10.212.10.188).

    Execute:

    root@bng-sbr-perfm3000-3: ./sbrd stop ssr

    Stopping ssr data processes

  3. Log in to a M node and execute ./sbrd status. Verify that the first D node is disconnected from the other nodes.

    Execute:

    root@bng-mars: ./sbrd status

  4. Log in to the second data node and execute ./sbrd status, and verify that it is still connected to the cluster. The following is an example of what the status of the second D node should look like:

  5. Log in to the first data node and then move your original installation to a different location to prevent the removal of files. Be sure not to move the /opt/JNPRshare directory, which should contain your original cluster configuration files. For example, if you previously installed Release 7.6.0, execute the following as root:

    1. Navigate to the opt directory: root@bng-sbr-perfm3000-3: cd /opt.

    2. Create a new directory: root@bng-sbr-perfm3000-3: mkdir 760.

    3. Move the configuration files to the new directory: root@bng-sbr-perfm3000-3: mv JNPRsbr JNPRmysql JNPRmysqld JNPRhadm JNPRndb_mgmd JNPRndbd 760.

    4. Remove the old software package: root@bng-sbr-perfm3000-3: pkgrm JNPRsbr.

  6. Navigate to the /opt/tmp directory and install the new SBRC software package.

    Execute:

    root@bng-sbr-perfm3000-3: pkgadd -d.

  7. Enter 1 to select the JNPRsbr.pkg.

    The script resumes.

  8. Confirm the installation directory.

    Depending on the system configuration, the script prompts you to create the /opt/JNPRsbr directory if it does not exist, overwrite an already-extracted package, or any of several other prompts.

    Answer the prompt appropriately (or change the extraction path if necessary) so that the script can proceed.

    To accept the default directory as a target, enter y.

    The script resumes.

  9. Enter y to confirm that you want to continue to install the package.

Once the system displays the message Installation of <JNPRsbr> was successful, the installation of the new software is complete on the first data node in the cluster. Once the software installation is completed, configure the new software on the data node and start the SSR process in the data node before repeating the installation steps on the other data nodes in the cluster. For more information about the software configuration, see Configuring the New Software on the Data Node.

Installing Only the Latest MySQL Package

Installing Only the Latest MySQL Package

To install the new software on the data node by copying only the latest MySQl package:

Caution

During your maintenance window, execute the following steps. If you run out of time during the maintenance window, it is possible to upgrade only some of the data nodes. This partially upgraded cluster has not been extensively tested by Juniper Networks and is not supported, so every attempt must be made to upgrade the entire cluster. It is recommended to install the new software by copying the complete SBR release distribution package (see Installing the Complete SBR Release Distribution Package).

Note

You must repeat this procedure on every D node in the cluster, one at a time.

  1. Verify that the /opt/JNPRmysql directory contains the directories install and mysql-cluster-advanced-7.4.15-solaris10-sparc-64bit. If these directories are not present, your installation is using non-default locations for these files. Locate these files and execute the steps in the following procedure in the same directory that contains these files.

  2. Verify that /opt/JNPRsbr/radius contains the sbrd file. If this file is not present, your installation is using a non-default location for the SBRC software distribution package. You need to locate the sbrd file and adjust the instructions below with the correct path to the file.

  3. Change to the directory located in Step 1, (usually, /opt/JNPRmysql) cd /opt/JNPRmysql.

  4. Unzip and un-tar the new MySQL package:

    Execute:

    gunzip -c /opt/tmp/JNPRsbr.pkg/reloc/radius/install/ndb/bin/*mysqlcluster*.tar.gz \ | tar -xf –

  5. Log in to the first D node as root, and stop the SSR process. In the following example, the D node is called bng-sbr-perfm3000-3.

    Execute:

    root@bng-sbr-perfm3000-3: ./sbrd stop ssr

    Stopping ssr data processes

  6. Check the status of the D node.

    Execute:

    root@bng-sbr-perfm3000-3: ./sbrd status

  7. Log in to the SM node and execute ./sbrd status, and verify that the D node is disconnected from the other nodes. The following is an example of what the status should look like:

  8. Remove the existing symbolic link and create a new symbolic link for the newly extracted MySQL package. For example:

    Execute:

    rm install

    ln -s /opt/760/mysql-cluster-advanced-7.4.15-solaris10-sparc-64bit install

  9. Start the SSR process on the data node.

    Execute:

    root@bng-sbr-perfm3000-3: /opt/JNPRsbr/radius/sbrd start ssr

  10. Check the status of the upgraded data node.

    Execute:

    root@bng-sbr-perfm3000-3: ./sbrd status

    Examine the display. The data node remains in a ‘starting’ state until it has copied all the data from the primary node with which it is paired. This process may take 30 minutes for a typical load. When the data node has completed, it will display as being reconnected to the cluster, as in the above example.

    Caution

    Once the data node finishes ‘starting’ and is shown in the same state as the other data nodes in the cluster, the upgrade of this data node is complete and the cluster is once again in a highly available state. It is imperative not to begin upgrading the next data node until this node has re-synchronized in this manner. Otherwise, it is possible to lose the cluster database contents and experience an outage.

  11. While waiting for the data node to complete the starting process and synchronize, monitor the connection state of the SM nodes in the cluster by repeating the ./sbrd status command. You should see a similar display, as follows:

    If the SM nodes suddenly disconnect, this is a symptom of a serious error in the SSR cluster that can occur under rare circumstances. If this occurs, you must manually restart the affected node by executing /opt/JNPRsbr/radius/sbrd restart radius to avoid an outage.

  12. Repeat this upgrade procedure until each data node has been upgraded.

Note

Software configuration is not required if you upgrade the software in D node by installing the MySQL package.

Configuring the New Software on the Data Node

Configuring the New Software on the Data Node

Configure the new SBRC software on the D nodes by running the configuration script.

Note

You must repeat this procedure on every data node in the cluster.

  1. As root, navigate to the directory where you installed the Steel-Belted Radius Carrier package.

    For example, navigate to the radius/install subdirectory.

    Execute:

    root@bng-sbr-perfm3000-3: cd /opt/JNPRsbr/radius/install/

  2. Execute the configuration script to configure the SBRC software.

    Execute:

    root@bng-sbr-perfm3000-3: ./configure

  3. Review and accept the Steel-Belted Radius Carrier license agreement.

    Press the spacebar to move from one page to the next. When you are prompted to accept the terms of the license agreement, enter y.

    Do you accept the terms in the license agreement? [n] y

  4. From the menu of configuration tasks, enter 3 to specify Configure Cluster Node.

  5. When prompted, enter the name of your existing cluster. In this example the cluster name is “6node_cluster”.

    You are prompted to verify whether you want to proceed, unless the script detects any unusual installation conditions (a pre-existing directory, for example). In some cases, you may be prompted to resolve or ignore them.

  6. The system reads the configuration files that you copied to the server and prompts you to change some settings to adapt them to this server.

    Enter a to accept the files.

  7. The system displays the following warning and prompts you to continue. Enter y to continue.

  8. When SSR is finished, you are returned to the main configuration menu.

    Enter q to quit.

  9. Start the SSR process on the data node.

    Execute:

    root@bng-sbr-perfm3000-3: /opt/JNPRsbr/radius/sbrd start ssr

  10. Check the status of the upgraded data node.

    Execute:

    root@bng-sbr-perfm3000-3: ./sbrd status

    Examine the display. The data node remains in a ‘starting’ state until it has copied all the data from the primary node with which it is paired. This process may take 30 minutes for a typical load. When the data node has completed, it will display as being reconnected to the cluster, as in the above example.

    Caution

    Once the data node finishes ‘starting’ and is shown in the same state as the other data nodes in the cluster, the upgrade of this data node is complete and the cluster is once again in a highly available state. It is imperative not to begin upgrading the next data node until this node has re-synchronized in this manner. Otherwise, it is possible to lose the cluster database contents and experience an outage.

  11. While waiting for the data node to complete the starting process and synchronize, monitor the connection state of the SM nodes in the cluster by repeating the ./sbrd status command. You should see a similar display, as follows:

    If the SM nodes suddenly disconnect, this is a symptom of a serious error in the SSR cluster that can occur under rare circumstances. If this occurs, you must manually restart the affected node by executing /opt/JNPRsbr/radius/sbrd restart radius to avoid an outage.

  12. Repeat this upgrade procedure until each data node has been upgraded; see Upgrading the Data (D) Nodes with the New Software.

    After all nodes are upgraded, you should see a similar output as follows: