Aws terraform deployment fails in zookeeper.conf task

**Describe the bug**
I am trying to deploy a Pulsar cluster in AWS following the documentation.
I manage to find my way to using `deploy_pulsar.yml`, but ansible fails in the `Add zookeeper.conf file` TASK with the message
```
TASK [Add zookeeper.conf file] **************************************************************************************************************************************************************************************************************
fatal: [zoo.kee.per.IP]: FAILED! => {"changed": false, "msg": "AnsibleUndefinedVariable: 'ansible.vars.hostvars.HostVarsVars object' has no attribute 'private_ip'"}
```

**To Reproduce**
Steps to reproduce the behavior:
1. Follow terraform instructions in https://pulsar.apache.org/docs/en/deploy-aws/
2. Skip the `Setup Disks` (does not work for me. i don't know what the authors wants us to do here)
3. Run Pulsar playbook
4. See error

**Expected behavior**
I am not an avid ansible user but i guess something like

`changed: [zoo.kee.per.IP]`
or
`ok: [zoo.kee.per.IP]`

**Screenshots**

**Desktop (please complete the following information):**
- Host OS: Windows 10
- Guest OS: Ubuntu 20 linux container

**Additional context**
I am running ansible from inside a containerized Ubuntu20 environment. My host OS is Windows 10.

Read the deploy-aws documentation and deploy_pulsar.yml, then inspect the Add zookeeper.conf file task and its hostvars/private_ip reference. Reproduce the failure with the documented Terraform and AWS setup, including the reported Setup Disks uncertainty, and verify that the playbook completes this task without AnsibleUndefinedVariable.

Written by the indexing model from the issue text.