Almalinux install times out because kickstart file is never downloaded.

[akulbe]

Code of Conduct

• I have read and agree to the project's Code of Conduct.

Project Version

main

VMware vSphere

8.0.0

HashiCorp Packer

1.8.5

HashiCorp Packer Plugin for VMware vSphere

1.1.1

Guest Operating System

AlmaLinux 9

Environment Details

vSphere & vCenter 8.0

running the packer stuff repo from a VM on this host

Description

Anaconda is failing to fetch the kickstart file, even though the HTTP server is starting.

Expected Behavior

Anaconda would get the kickstart file, AlmaLinux install would complete successfully as expected.

Actual Behavior

The KS file is never downloaded, and the installation times out.

Steps to Reproduce

Run set-envvars.sh, supply the following answers:

[akulbe@workspace packer-examples-for-vsphere]$ ./set-envvars.sh

> Set the vSphere credentials.
Enter the FQDN of your vCenter Server instance: vcsa.kulbe.cloud
Enter the username for the user account: [email protected]
Enter the password for the user account:
Skip vCenter Server instance SSL verification? (y/n): y

> Set the vSphere settings.
Enter the vSphere datacenter name: kulbe-dc
Enter the vSphere cluster name: kulbe-cluster
Enter the datastore name virtual machines: datastore2
Enter the network name: VM Network
Enter the folder name: Packer Templates
Enter the content library name: LVB
Enter the datastore name for .iso files: datastore1

> Set the common virtual machine settings.
Enter the virtual hardware version (recommended: 19): 20
Enable VMware Tools upgrade (recommended: true): true
Remove CD-ROMs (recommended: true): true

> Set the common template and content library settings.
Convert to a template (recommended: false): false
Export OVF Template to the content library (recommended: true): true
Destroy virtual machine artifact (recommended: true): true
Skip export to content library (recommended: false): false

> Set the common OVF export settings.
Export as an OVF artifact (recommended: false): false
Overwrite an existing OVF artifact (recommended: true): true

> Set the common boot and provisioning settings.
Set the provisioning datasource (recommended: http): http
Enter the IP address of the interface on this host: 172.20.252.250
Enter the starting HTTP port (recommended: 8000): 8000
Enter the ending HTTP port (recommended: 8099): 8099
Enter the IP wait timeout (recommended: 20m): 20m
Enter the virtual machine shutdown timeout (recommended: 15m): 15m

> Set proxy credentials.
Use SOCKS Proxy? (y/n): n

> Set the default user account credentials.
Enter the username for the account: aaron
Enter the password for the account:
Enter the SHA-512 encrypted password for the account:
Enter the key for the account:

> Set the Ansible credentials.
Enter the username for the account: ansible
Enter the key for the account:

> Set the RedHat Subscription Manager credentials.
Enter the username for the account: longviewbits
Enter the password for the account:

> Set the SUSE Customer Center credentials.
Enter the email for the account:
Enter the code for the account:

> Set the HCP Packer registry.
Enable the HCP Packer registry: no


> Setting the vSphere credentials...
> Setting the vSphere settings...
> Setting the common virtual machine settings...
> Setting the common template and content library settings...
> Setting the OVF export settings...
> Setting the common boot and provisioning settings...
> Setting the default account credentials...
> Setting the Ansible credentials...
> Setting the RedHat Subscription Manager credentials...

> Setting the SUSE Customer Center credentials...

> Setting the HCP Packer...

Display the environment variables? (y/n): n
[akulbe@workspace packer-examples-for-vsphere]$

Log Fragments and Files

No response

Screenshots

Additional Context

I realize this could be a configuration issue on my part, but I'm honestly not sure.

[tenthirtyam]

I’ve converted the issue to a discussion - all builds were tested in main yesterday.

This appears to be a configuration issue with the machine image being build not having access back to the Packer host’s HTTP server - you can can switch to disk-based mode per the README to test if the build completes.

[akulbe]

This is the part that I don't understand. The network is flat. A single subnet. So I'm not sure why it would be unreachable.

[tenthirtyam]

Assuming that you're Linux, have you configured IPTables?

[akulbe]

🤦🏻‍♂️ it helps when you reload the firewall.

Sorry for this. I figured it was something I missed.

On another note… THANK YOU for this work. 🙌🏻

[tenthirtyam]

That's great to hear - you're welcome!

Be sure to give us a GitHub ⭐!

Ryan Johnson
Senior Staff Solutions Architect | Product Engineering @ VMware, Inc.