Puppet 6 client exits on start when it cannot connect to puppetserver

Puppet Version: 6.4.2
Puppet Server Version: 6.3.0
OS Name/Version: CentOS 7.6

In previous versions of puppet, I would spin up instances in AWS, regardless of the puppetservere being available, and have runinterval at 20 seconds. Whenever the base packages were installed, I'd update the puppet configuration to a runinterval of every 30 minutes.

Since I've started version 6 clients, this "trick" no longer works. The puppet client will try to connect to the puppetserver when it starts, and if it doesn't it will exit. That's quite disastrous for my deployment strategy.

Steps to reproduce:

Pre-requisites: Have the above puppet server and client versions on a single machine. Have puppetserver and client stopped.

1. Start puppet client.

Desired Behavior:

A. puppet client starts, regardless of a connection is possible with the puppetserver. It will re-try each runinterval.

Actual Behavior:

• puppet client exits with below message in /var/log/messages:

  May 22 15:08:05 bas-t-1 systemd: Started Puppet agent.

  May 22 15:08:07 bas-t-1 puppet-agent[12580]: Could not run: Failed to open TCP connection to bas-t-1.t.dam.loc:8140 (Connection refused - connect(2) for "bas-t-1.t.dam.loc" port 8140)

  May 22 15:08:07 bas-t-1 systemd: puppet.service: main process exited, code=exited, status=1/FAILURE

  May 22 15:08:07 bas-t-1 systemd: Unit puppet.service entered failed state.

  May 22 15:08:07 bas-t-1 systemd: puppet.service failed.

  puppet.conf:

  [main]

  server = bas-t-1.t.dam.loc

  environment = production

  runinterval = 20s

  hiera_config = /etc/puppetlabs/puppet/hiera.yaml

  [master]

  dns_alt_names = bas-t-1.t.dam.loc