when a packer server has more then one NIC, Packer can often
select the wrong http server using {{ .HTTPIP }}. We extend the
variables file to include an http_server_address which can be set
to the IP address of the correct NIC. This configuration propagates down
to the http_bind_address directive of the builder and the boot_command
entered on the guest VM. Adding 0.0.0.0 as a reasonable default as it
preserves comptability with packers default behavior.
This commit was required to get Packer to build the ubuntu guest
successfully in our environment
Additional Reading:
https://github.com/hashicorp/packer/issues/9973https://www.packer.io/docs/builders/vmware/vsphere-iso#http-directory-configuration
Since Terraform 0.13.x there is no need to build the provider or download it manually. The file versions.tf was added to specify the provider name in the Terraform Registry. The instructions in the README file have been updated to reflect this change.
Update ESXi provider to josenk from hashicorp/esxi
```
DetectionLab/ESXi terraform init && terraform apply
Initializing the backend...
Initializing provider plugins...
- Finding latest version of hashicorp/esxi...
Error: Failed to install provider
Error while installing hashicorp/esxi: provider registry registry.terraform.io
does not have a provider named registry.terraform.io/hashicorp/esxi
```
The syntax "sudo ifconfig up eth1" was throwing an error in the current version of Ubuntu. The "up" option needs to come after the interface name for it to work.
Added ssh_timeout of 10000 seconds (166.6667 minutes). This will keep packer from destroying the VM before it's even done with the setup process. The time can be adjusted but this helped me in the setup process.
