Basic pre-requisites for setting up a cluster
Setup the hostname
Setting up vansetup.conf
Executing vansetup.py
Setting up "Analytics cluster" on the Director
Generating certificate on the Director (if applicable)
Syncing Director certificate with Analytics
Setting up a log-forwarder (if applicable)
Configuring the hosts under "settings"
Editing the data retention/configuration
Check for the common cron jobs
Verify ADC configuration on the Controllers (if applicable)
Verify LEF connection on the Branches
Basic pre-requisites for setting up a cluster
The cluster nodes can be brought up as Baremetal or VMs, the below link provides aspects on scalability


Set up the management, listener/replication and collector interfaces with the appropriate addressing (refer to the Analytics architecture KB)
Setting up the hostname
The initial state of a fresh analytics node is as below, versa-analytics-driver would be in stopped state and the default hostname would be versa-analytics

To setup the hostname you can edit the /etc/hosts file with the appropriate hostname to address mapping (the hostname can be mapped to the listen address or rpc address)

Edit the /etc/hostname file and set the appropriate hostname

Reboot the node for the hostname to take effect

Setting up the vansetup.conf
The vansetup.conf file present under /opt/versa/scripts/van-scripts is the file that is read by vansetup.py towards setting up the analytic/search personality and installing the required services.
The cluster-name would need to be consistent among all the nodes in the cluster. Also, ensure that the correct personality is configured. along with the rpc (management) and listen (replication/internal) address.
The "seeds" value on the first node that is being brought up should be set to the self listener address
The "seeds" value on the rest of the nodes, that are brought up, should be set to "<listen address of existing node>,<self listen-address>"
Example configuration on the first node that's being brought up in the cluster is as below

On the rest of the nodes being brought up in the cluster the seeds value would be as below "<listen address of an existing node>,<self's listen address>"
