Configuring and running the built-in balancer on a cluster node

23 May 2024

ID 243044

Only applicable when integrating with multiple servers of КАТА 5.0 or a later version.

To configure the built-in balancer on a cluster node:

  1. Connect to the cluster node over SSH to gain access to the Technical Support Mode.
  2. Rename the basic configuration file of the balancer with the following command:

    mv /etc/haproxy/haproxy.cfg /etc/haproxy/haproxy.cfg.orig

  3. Place the previously prepared configuration haproxy.cfg configuration file in the /etc/haproxy directory and set access rights for it using the following commands:

    chown root:root /etc/haproxy/haproxy.cfg

    chmod 640 /etc/haproxy/haproxy.cfg

  4. Create the /etc/rsyslog.d/haproxy.conf file with the following content:

    $ModLoad imudp

    $UDPServerRun 514

    $UDPServerAddress 127.0.0.1

    if $syslogfacility-text == 'local6' then /var/log/haproxy.log

    if $syslogfacility-text == 'local6' then stop

  5. Create the /var/log/haproxy.log file and set access rights for it using the following commands:

    touch /var/log/haproxy.log

    chown root:adm /var/log/haproxy.log

    chmod 640 /var/log/haproxy.log

  6. If necessary, edit the log rotation settings file for the log balancer (/etc/logrotate.d/haproxy).

    By default, entries are rotated daily. Entries are kept for 10 days.

  7. Restart the system log service:

    systemctl restart rsyslog

  8. Start the built-in balancer service:

    systemctl start haproxy

  9. Allow running the built-in balancer service automatically:

    systemctl enable haproxy

  10. Check the status of the built-in balancer service:

    systemctl status haproxy

    The status must be running.

  11. Check the presence of entries in the built-in balancer log:

    tail /var/log/haproxy.log

The built-in balancer is configured and started on the Kaspersky Secure Mail Gateway cluster node.

Did you find this article helpful?
What can we do better?
Thank you for your feedback! You're helping us improve.
Thank you for your feedback! You're helping us improve.