====== Primary and backup configuration ======

How to setup Trisul Hub and Probe in a High Availability configuration.  This guide describes how a backup probe and hub can be setup to take over in case of a primary failure.

===== Setup =====

You need an identically sized backup installation of Trisul that can takeover the load in the event of a primary failure.

  * Download the replication scripts [[https://trisul.org/download/sync.sh|sync.sh]] and [[https://trisul.org/download/host_alive.sh|host_alive.sh]]

{{:admin:haprimarybackup.png?400|}}

===== Configure SSH between the primary and backup nodes =====

On both the primary and backup node use [[https://www.ssh.com/ssh/copy-id|ssh_copy_id to allow login]] without an interactive password. 



===== Data replication =====

The primary hub and probe nodes contains the database and configuration. We use a script based on RSYNC to replicate this database incrementally over to the backup node. 

  - Download the script sync.sh
  - Edit the script to enter the IP Address of the Backup script
  - In a SCREEN session run the backup script

**On the primary node**

<code>
screen
./rsync.sh
</code>


===== Heartbeat check =====

The backup node needs to keep checking the liveness of the primary node using a heartbeat.  


**On the backup node**

<code>
screen
./host_alive.sh
</code>

 
===== Takeover =====

The keepalive uses a PING to check the reachability of the primary node. When a configured number of PINGs (default 6 PING failures) fail, the backup starts the Trisul HUB ad PROBE processes and this takes over the primary function.


===== Recovery =====

The recovery process is  not automatic.  After the primary node is restored, you stop the Trisul processes on the backup and start the process on the primary node.