I have been using Vagrant for few months. Also I have started using Virtual box without Vagrant with sharing folders from host OS and it works well so far.
My question is why is Vagrant needed if Virtual Box can share folders without Vagrant?
Maybe I don't use Vagrant's other features so I don't need them so far.
Just a few things, but will probably not be exhaustive:
Vagrant operates multiple provider, not only Virtualbox, you can run vagrant with Virtualbox, VMWare, HyperV ... (docker as well even if there are different pros and cons) so vagrant abstracts this for you
vagrant can setup shared folders as you experienced, but can easily set different shared folder types (nfs, rsync ..) depending your setup
vagrant will manage all the networking of the VM, if you need a static IP, it will associate hostname and static IP and setup all the routes for you.
vagrant works well with many provisioning tools (puppet, chef, ansible ...) so you can easily create and re-created multiple times the same environment
On top of all this, why is vagrant good to use ? In team. If you work in team, you share a vagrant file (just a ruby script file) and "magically" all your team members share the same environment to work.
Related
I currently have a windows 10 which I like for everything, except when it comes to development. So whenever I need to work, I run an Ubuntu virtual box machine and I code there.
I recently started working with a team that uses Vagrant boxes (For development environments) and Ansible (For provisioning). My question is, is it possible to run a Vagrant box inside my Ubuntu virtual machine to provision it with Ansible? (Since it is only available on Linux distros).
I realize that this is a bit confusing since I am running a virtual machine inside another. Is there any way I can go around this?
Vagrant is able to run on windows, you can set a nice directory sharing and you can access that directory from your other VM. If provision is part of the Vagrantfile, you don't have to worry whether this will work in windows or not because most of the provisioners run inside the created machine already.
For Ansible, you should use Ansible Local provisioner. That will let it run on guest machine completely and you won't need Ansible on the host machine.
The majority of developers on our project use Windows but our servers are Linux and provisioned with ansible.
In order to test ansible locally we are using Vagrant to create VMs to simulate the servers but we also need to create a VM for ansible as it doesn't support Windows as a control machine.
Some developers use Linux or Macs so they don't need this extra VM.
How do I configure and run Vagrant so that it only provisions this extra VM if I ask it to?
Or alternatively, how do I prevent it provisioning this VM if I tell it not to?
I think some excerpts from the docs can help you:
Other commands, such as vagrant up, operate on every machine by default. So if you ran vagrant up, Vagrant would bring up both the web and DB machine. You could also optionally be specific and say vagrant up web or vagrant up db.
Additionally, you can specify a regular expression for matching only certain machines. This is useful in some cases where you specify many similar machines, for example if you are testing a distributed service you may have a leader machine as well as a follower0, follower1, follower2, etc. If you want to bring up all the followers but not the leader, you can just do vagrant up /follower[0-9]/. If Vagrant sees a machine name within forward slashes, it assumes you are using a regular expression.
Also, a "primary" machine:
You can also specify a primary machine. The primary machine will be the default machine used when a specific machine in a multi-machine environment is not specified.
To specify a default machine, just mark it primary when defining it. Only one primary machine may be specified.
You can define one VM as primary, so vagrant up will provision only that machine, and you will have to manually pass all other VM names to provision them.
Or you can write some shell script like up.bat and up.sh and sequentially call vagrant up VMname only for those VMs that are required for that OS.
Hadoop & Vagrant: If I set up a new VM using these two packages, is the VM available for starting with VirtualBox or do I always have to use vagrant up?
When you first create the VM, you will need to use vagrant up since VirtualBox has no ability to read a Vagrantfile. Once your VM is up and running, you could stop and start it using VirtualBox. However, Vagrant can perform the same operations as VirtualBox and has a lot of advantages on top of that.
First let's discuss how Vagrant can do what VirtualBox can do. You can use vagrant suspend to put the VM in a saved state, vagrant resume to start it again, vagrant halt to power it off, and vagrant destroy to delete it. You can also use settings in Vagrant's VirtualBox provider to modify your VM's memory, CPUs, and more. You can also change the network settings and synced folders. And this gives you the advantage of defining all of these settings in code which can be checked into a revision control system such as Git.
On top of all this, Vagrant has support for provisioners. These range from something as simple as shell scripts to full blown configuration management tools such as Puppet, Chef, Ansible, and Salt. Just like you can define your VM's VirtualBox settings in code, you can define the entire creation process including installing packages, managing services, and customizing files. And you can make changes at any time and apply them using vagrant provision.
So getting back to your question... yes, you can control your VM directly with VirtualBox, but you'd be missing out on the rich feature set of Vagrant.
I currently have a Vagrant vm with a typical web setup (apache, php, etc). The actual web repo is checked out onto the my local machine, and accessed through synced folders and forwarded apache ports (like http://127.0.0.1:4567/web). The question is, how do I make this work for a team, as opposed to just me? Since you have to sync folders, how is it possible for multiple devs to connect to one VM and sync folders?
I initially thought sharing the Vagrantfile would be all that is needed, but the Vagrantfile is agnostic to all apt-get commands and apache/php/node etc setup. I don't want devs to have to worry about running those.
This is what Vagrant provisioning is for.
Vagrant allows you to automatically alter configuration of a box or install additional software packages as a part of the first vagrant up process. Moreover, you can provision your system with various configuration management systems: Chef, Puppet, Ansible, CFEngine. Or you can simply use shell scripts.
Sitepoint offers a good tutorial on provisioning Ubuntu 14.04 box with bash scripts to install nginx, PHP-FPM and MySQL. Read it. Later you might want to move to configuration management systems.
Take a look at Phansible project which generates Ansible provisionings for PHP-based projects. Puppet provisioning for PHP-projects can be generated via PuPHPet. They will give you an idea how to use configuration management with Vagrant and might be used as s template for your Vagrantfile.
I always try to craft my Vagrantfiles carefully (provisioning, synced folder, for example, maps host's ./src/public onto /var/www/html/ on guest machine, port forwards and so on) so I can put them into project's root under version control. Later when my teammate clones project's repo on his machine he can issue vagrant up right away and get a fully functional development environment.
The only problem I haven't solved yet is updating existing VMs when new dependencies are added to projects. Now we update provisioning scripts and if developers encounter errors they manually reprovision their VMs by using vagrant reload --provision.
P.S. I asked a question regarding this problem
I have a Vagrant set up with 3 virtual machines. Each machine has its own shell script for provisioning.
Now I would like to share the exact same status of my set up with somebody else. Since the provisioning procedure takes really (!!) long for each machine, I hope there is another solution.
Ideally I would be able to save each machine as it is in one file, which the other person then could import into Virtualbox. Is there a way to do that?
If I understand you correctly you would like to make a Vagrant base box from provisioned by Vagrant VMs. This is not recommended way to go. How you can approach this is:
Create new VM manually with required OS in the VBox.
Adjust it so Vagrant can connect to it as described here and here.
Provision it using your shell scripts.
Install all the things you would find useful to have on this VM.
Use Vagrant to package it as a base box as described here.
After packaging it with Vagrant you will get a Vagrant base box file with .box extension. You can then pass this to your team mates (usb, network share, ftp etc.) and they can add it to their Vagrant installation and use it. Whenever they will do Vagrant up they will get fully provisioned VM in VBox with all the stuff you have packaged to it. Vagrant also gives you versioning capabilities. If properly configured whenever you will create new version of base box everybody who is using it will be notified and would be able to download and use new version of your box.
Hope I understood your problem correctly and this will help to solve it.