Enabling or disabling high availability (HA) - AWS Elemental Conductor File

Enabling or disabling high availability (HA)

Disable high availability prior to performing any changes on the Conductor nodes.

To disable high availability on a conductor:

  1. From the Linux prompt, log in to the conductor node with the elemental user credentials.

  2. Enter the following command to configure the primary Conductor node for HA.

    [elemental@hostname ~]$ sudo /opt/elemental_se/.support_utils/dbrepl disable

  3. Restart the service using the following commands.

    [elemental@hostname ~]$ sudo systemctl restart postgresql-15 
[elemental@hostname ~]$ sudo systemctl restart elemental_se

  4. Enter the following command to verify that Conductor high availability is disabled.

    [elemental@hostname ~]$ tail -F /opt/elemental_se/web/log/conductor.output

    The conductor.output log starts to scroll on the screen and shows messages as they are occurring. Watch for the following INFO lines on the primary Conductor node.

    CONDUCTOR: Initializing environment
I, [2024-06-04T23:07:54.439807 #131824]  INFO -- : HA environment not enabled
[2024-06-04 23:08:00 UTC SERVICE]: Elemental Conductor File 2.18.x.x

  5. Enter Ctrl+C to exit the tail command.

  6. Enter the following command, where <day> is today (the day you are upgrading), typed with an initial capital letter: Mon, Tue, Wed, Thu, Fri, Sat, Sun:

    [elemental@hostname ~]$ sudo tail -F /data/pgsql/logs/postgresql-<day>.log

  7. Confirm that you see this line.

    database system is ready to accept connections.

    If the elemental_se or postgres process has already started when you starting tailing the logs, you might not see the ready to accept connections message. Instead, you could see rejects connection for host messages until you upgrade the worker nodes.

  8. Enter Ctrl+C to exit the tail command.

You must re-enable high availability on the Conductor nodes, to put the nodes back into a redundant configuration. If you have only one Conductor File node, skip this step.

To enable high availability on the primary Conductor

  1. From the Linux prompt, log in to the primary Conductor node with the elemental user credentials.

  2. Enter the following command to configure the primary Conductor node for high availability. Where <dbrepl_config_file_name> is your dbrepl_config.yml file.

    [elemental@hostname ~]$ sudo /opt/elemental_se/.support_utils/dbrepl configure <dbrepl_config_file_name> primary

  3. Restart the elemental_se service.

    [elemental@hostname ~] sudo systemctl restart elemental_se

  4. Enter the following command to verify that the service is running.

    [elemental@hostname ~] tail -F /opt/elemental_se/web/log/conductor.output

    The conductor.output log starts to scroll on the screen and shows messages as they are occurring. Watch for the following three INFO lines on the primary Conductor node.

    CONDUCTOR: Initializing environment
I, [2024-06-04T20:58:01.073149 #9844] INFO -- : Configuring the HA environment
I, [2024-06-04T20:58:03.170375 #9844] INFO -- : Preparing database as replication master
...
[2024-06-04 20:58:09 UTC SERVICE]: Elemental Conductor File 2.18.x.x

  5. Enter Ctrl+C to exit the tail command.

  6. Enter the following command, where <day> is today (the day you are upgrading), typed with an initial capital letter: Mon, Tue, Wed, Thu, Fri, Sat, Suns.

    [elemental@hostname ~]$ sudo -s
[elemental@hostname ~]$ cd /data/pgsql/logs
[elemental@hostname ~]$ tail -F postgresql-<day>.log

  7. Confirm that you see this line.

    database system is ready to accept connections.

    If the elemental_se or postgres process has already started when you starting tailing the logs, you might not see the ready to accept connections message. Instead, you could see rejects connection for host messages until you upgrade the worker nodes.

  8. Enter Ctrl+C to exit the tail command.

  9. Type the following command to exit the session as the sudo user.

    [elemental@hostname ~]$ exit

On the secondary Conductor

Repeat high availability steps on the secondary Conductor but use the following command instead.

[elemental@hostname ~]$ sudo /opt/elemental_se/.support_utils/dbrepl configure <dbrepl_config_file_name> secondary