Ubuntu Vmware Image

[Stefanie Mordaunt]

Ihave been stuck on a netbook since last year due to budget constraints, but finally managed to get a decent laptop from a very generous person. Now my ubuntu installation in the netbook has all my development tools, libraries, personal mediawiki, other servers and things that I'm only reminded when the command line complains. I can always ssh into my netbook, but don't want to carry both the computers all the time. Is it possible to create an image from my netbook Ubuntu and use it out of the box in a Vmware player in the new laptop?

If I use a new ssh-key ( generate new from the console ), or use an older key, once it's booted up, these keys have no effect on establishing an SSH connection. It will prompt for password, and this is not due to permissions issue on the host key side just fyi. The old password on the system doesn't work with any combination of the legitimate old user, or if it was replaced ( not sure if it will do this ) with "ubuntu/root/ec2-user".

I tried SSHing from an older machine just in-case since the 14.04 is on openssh 6.6x but has no effect. Would someone possibly point me in the right direction to resolve this as quickly as possible? Or is the only solution to fix the original image pre-AMI, and do some specific adjustments?

I haven't used the VM import feature for a few years, but according to documentation, it shouldn't change the configuration of users present on the system, and it doesn't create the ec2-user (or "ubuntu" in the case of AWS's standard Ubuntu AMIs): -import/latest/userguide/prepare-vm-image.html#prepare-vm-image-linux

To be sure, you are able to log on to the system with a public key at the source, but the exact same key supplied to the exact same user as before is no longer accepted after launching the image as an EC2 instance?

If you want a quick workaround instead of troubleshooting the import tool, you could launch a brand new, temporary instance in the same AZ from a standard AWS AMI, detach the root volume from the imported instance, attach it to the temporary instance, and inspect and correct the credentials configuration there. Then just reattach the volume to the original instance and boot it up.

Specific Deep Learning VM Images images are available to suit your choice offramework and processor. There are currently imagessupporting TensorFlow, PyTorch, and generic high-performancecomputing, with versions for both CPU-only and GPU-enabled workflows.To find the image that you want, see the table below.

Choose a Deep Learning VM image family based on the frameworkand processor that you need.The following table lists the most recent versions of image families,organized by framework type.To get the most recent version of an image, create an instanceby referencing an image family with latest in the name.If you need a specific framework version, skip to Supported frameworkversions.

For most frameworks, Debian 11 is the default OS. Ubuntu 22.04images are available for some frameworks.They are denoted by the -ubuntu-2204suffixes in the image family name (see Listing all availableversions).Debian 10 and Debian 9 images have been deprecated.

TensorFlow Enterprise image families provide a Google Cloudoptimized distribution of TensorFlow.For more information about TensorFlow Enterprise including which versionsare supported, seeTensorFlow Enterprise overview.

Some Deep Learning VM image families are experimental, as indicatedby the table of image families. Experimental images aresupported on a best-effort basis, and may not receiverefreshes on each new release of the framework.

You can reuse the same image even if the latest image is newer. This can beuseful, for instance, if you are trying to create a cluster and youwant to ensure that any images that are used to create new instances arealways the same. You should not use the name of the image family in thissituation because, if the latest image is updated, you'll have different imageson some instances in your cluster.

To find out the exact name of the latest image, use the following command inthe Google Cloud CLI with your preferred terminal or inCloud Shell. ReplaceIMAGE_FAMILY with the image family name for which you wantto find out the latest version number.

Deep Learning VM supports each framework version based on a schedule tominimize security vulnerabilities. Review the Deep Learning VM frameworksupport policy to understand theimplications of the end-of-support and end-of-availability dates.

Image families are named in the formatFRAMEWORK-VERSION-CUDA_VERSION(-experimental),where FRAMEWORK is the target library,VERSION is the framework version, andCUDA_VERSION is the version of the CUDA stack, if present.

Except as otherwise noted, the content of this page is licensed under the Creative Commons Attribution 4.0 License, and code samples are licensed under the Apache 2.0 License. For details, see the Google Developers Site Policies. Java is a registered trademark of Oracle and/or its affiliates.

I am a HUGE fan of HashiCorp Packer and I have been using it for a number of years across many different projects including the VMware Event Broker Appliance (VEBA) solution. While it can certainly feel daunting at first, the same can be said for just about anything new, I typically point folks over to Ryan Johnson's fantastic Packer Examples for VMware vSphere project as a starting point, where you can find working Packer examples across a number of popular OS distributions for both Windows and Linux.

Most recently, I was helping out a few colleagues who was interested in automating the build of an Ubuntu Desktop image that could then be exported to an OVF/OVA. Of course, my recommendation was for them take a look at Ryan's project and they should be able to augment the existing Ubuntu Server 22.04 example. Interestingly enough, while I always recommend Ryan's Packer example repo, I have not personally used it myself and this is primarily due to the existing customization I have in my Packer builds which includes the use of custom OVF properties, which you can read more about HERE, HERE AND HERE.

Ryan's project is extremely comprehensive and while things should just work if you use the default builds, but if you wish to make tweaks, I can certainly understand that you could feel overwhelmed, which is exactly how I felt when trying to figure out how to augment the existing Ubuntu Server 22.04 build.

While I do have experience in using Packer, it did take me a few attempts as I ran into some setup issues on my macOS system and just ended up deploying an Ubuntu 22.04 VM to then use as my build host. The required change to go from an Ubuntu Server to Ubuntu Desktop was minimal, you do need to understand the project layout and ultimately how the repo has been setup, which includes the use of Ansible Packer Provisioner, which was not something I had used before.

I wanted to put together this blog post, not only as a reference for myself but also for anyone who wants to start using Packer and Ryan's awesome repo but need a bit more guidance if you intend to perform further customization.

Step 1 - Download and install the latest Ubuntu Server 22.04 ( -22.04.2-live-server-amd64.iso) as a VM. While you can certainly run the Packer build from your desktop system, depending on how it has been setup, you might still run into issues and having a dedicated build host is not necessary a bad thing.

Step 5 - Create a new build configuration, which allows you to isolate each build separately including credentials and build environments. In the example, I have named this jammy_desktop, but you can use any name as it simply creates a directory with the Packer configuration files that you will need to edit.

Step 6 - Depending on when you clone Ryan's repo, you may want to use newer versions of the OS ISO images and that requires you to update the Packer configuration file to include the directory where the ISO image will be stored on your vSphere Datastore but also the filename and SHA256 hash. In my example, I am using the latest Ubuntu Server 22.04.2 image and I have tweaked the filename to match the default value but the directory and the SHA26 hash is a different value. You can use sha256sum utility to generate the required has.

Step 7 - Since we want an Ubuntu Desktop image rather than the default Ubuntu Server, which does not include a graphical desktop, we will need to adjust the cloud-init user-data configuration to include the additional ubuntu-desktop package which will give us the desktop experience.

Note: If you need to make other OS customization, you can certainly apply the changes here or look at using the Ansible Packer Provisioner which is also used within the project. Additional research should be done if you intend to use Ansible rather than native cloud-init interface which Ubuntu supports.

Additionally, we also need to provide a password for the build account including the SHA512 hash of the password. To generate the hash, you can run the following command and then specify the password of your choice and it will output the SHA512 hash

Step 11 - Now we need to edit common.pkrvars.hcl file and specify the vSphere Datastore in which your ISO images will be stored by editing the common_iso_datastore variable. Ryan's repo uses both the Content Library and OVF Providers to allow multiple output of the final image. If you do not have or intend to output the Ubuntu image into an existing vSphere Content Library, then make sure you configure the common_content_library_skip_export variable to true or else you will hit an error at the very end of the build.

Step 12 - The last configuration file you will need to edit is the vsphere.pkrvars.hcl file which contains information about your vSphere environment including vCenter Server, credentials and resources Packer will use. This should be pretty straight forward and if you are using the default self-signed TLS certificate for vCenter Server, make sure you set the vsphere_insecure_connection variable to true.

3a8082e126