Since Ubuntu 15.10 Desktop has reached the end of its support life-cycle, I have updated the prototype to use Ubuntu 16.04 Desktop. I recommend re-provisioning the prototype VM as follows:
Re-provisioning the Prototype VM
0. If you had stored any information on the prototype VM that you wish to keep, first copy it off the VM so that it is not lost. The easiest way is to create a new folder within the /vagrant shared folder and put the data in there. In general, avoid keeping data on the prototype VM, since everything on it should be dynamically generated so that it’s easy to re-provision the VM as needed.
1. Make sure the VM is shut down. On the host machine (your desktop or laptop computer) switch to the folder where you have checked out the analytics-prototype git reposity and then run
vagrant destroy
2. The above command will delete the prototype VM from VirtualBox, however, it will not remove the base box. If you wish to free the disk space used by the older base box, you may run this command:
vagrant box remove kevinmellott91/ubuntu-15.10-desktop-amd64
3. Next, use your Git tool of choice to pull down the latest changes to the analytics-prototype git repository.
4. Now you can provision the new prototype VM with:
vagrant up
5. This will pull down the new base box and run the provisioning script. Be sure to allow the provisioning script to fully complete before interacting with the VM. IMPORTANT: It’s best to shut down the VM after first-time provisioning to allow all of the installed software to settle. Then start it a second time with the vagrant up command.
Notes About this Update
I tested several of the Ubuntu 16.04 Desktop base boxes available on Vagrant Atlas but found issues with each one I tested. As a result, I decided to construct a new base box. Going forward, the prototype will use jamesthinkbox/ubuntu-16.04-desktop-amd64 as the base box (as can be seen inside the Vagrantfile). In general, Ubuntu 16.04 Desktop appears to be much more stable than 15.10, and the first-time provisioning process goes much more smoothly with no visible errors.
To date I have not heard back from the folks at SlamData, so I have removed it from the provisioning scripts until I can get it sorted. As a result, none of the visualization examples will work at this time. However, it is still possible to run the generator and inspect the resulting analytics data using Robomongo. The provisioning script now installs Elasticsearch and Grafana, which should form the basis for forthcoming visualization examples.
If you have any questions or difficulties with the re-provisioning process, don’t hesitate to post here.