I am using Vagrant to house my Rails application. Unfortunately I am hitting a blocker in that using NFS to mount my synced folder results in errors with anything reading from or writing to the synced folder; Resque, Solr and even bundler.
Using bundle install --path vendor/bundle, for instance, results in:
Bundler::GemspecError: Could not read gem at /var/applications/beatroot/vendor/bundle/ruby/2.1.0/cache/sunspot_rails-2.1.1.gem. It may be corrupted.
Reverting to using a non-NFS synced_folder resolves the issue, leading me to believe it's a permissions problem. Unfortunately the performance loss necessitates using NFS.
Has anyone else encountered this? Is there a solution?
The host is OS X 10.9.2 and the guest is Ubuntu Precise 64.
My vagrant file looks like this:
VAGRANTFILE_API_VERSION = "2"
Vagrant.configure(VAGRANTFILE_API_VERSION) do |config|
config.vm.box = "precise64"
config.vm.box_url = "http://files.vagrantup.com/precise64.box"
config.vm.provider :virtualbox do |virtualbox|
virtualbox.name = "beatroot-development"
virtualbox.memory = 3072
virtualbox.cpus = 4
end
# Provision dependencies from shell
config.vm.provision :sprinkle do |sprinkle|
sprinkle.script = "provisioning/bootstrap_vagrant.rb"
sprinkle.cloud = true
sprinkle.verbose = true
end
# Forward port to Unicorn
config.vm.network :forwarded_port, host: 5009, guest: 5009
# Create a private network to allow for NFS mounting
config.vm.network :private_network, ip: "192.168.50.50"
# Mount our app in the VM
# We use NFS for performance gain
config.vm.synced_folder "./", "/var/applications/beatroot", nfs: true
end
Related
I'm stuck again. I need to provision a multi-machine Environment - one VM for a Sinatra app and a second for its PostgreSQL DB.
So far, I've managed to get the Sinatra app up and running in the ubuntu/xenial64 box but the provisioning "breaks" when it hits the configuration for the DB
Vagrant.configure("2") do |config|
config.vm.define "app" do |app|
# Use ubuntu/xenial64 as the virtual machine
app.vm.box = "ubuntu/xenial64"
# Use a private network to connect the VM to the local machine via an IP with an alias
app.vm.network "private_network", ip: "192.168.10.100"
app.hostsupdater.aliases = ["development.local"]
# sync the 'app' directory in the local directory to '/app' on the VM
app.vm.synced_folder "app", "/app"
# Use the provisioning script in envirnonment to provision the VM for a Sinatra environment
app.vm.provision "shell", path: "environment/app/provision.sh"
app.vm.provision "shell", inline: set_env({ DATABASE_URL: "postgresql://myapp:dbpass#localhost:15432/myapp" })
end
config.vm.define "db" do |db|
db.vm.box = "ubuntu/trusty64"
db.vm.host_name = "postgresql"
db.vm.network "private_network", ip: "10.0.2.15"
# db.vm.forward_port 8000, 8000
db.hostsupdater.aliases = ["database.local"]
# db.vm.share_folder "home", "/home/vagrant", ".", :create => true
db.vm.provision "shell", path: "environment/db/provision.sh", privileged: false
end
end
As you've probably guessed, I'm running an external provisioning script for the PG setup. The odd thing is I'm using the script recommended from Postgres' own site here.
In a separate location, I've git cloned that repo and followed the instructions and it works absolutely fine, creating a properly provisioned VM with PG installed.
However, I want to run a single vagrant up command and provisioning both the app and db correctly and have them speak to each other.
I'm (quite clearly) new to provisioning and DevOps as a whole, so would really appreciate some help.
I've uploaded my hilariously broken code here for you kind souls to look over if you feel so inclined.
Vagrant documentation on Multi-machines is quite thin and Google isn't being much help
Thanks!
I am using windows and putty to ssh to vagrant virtualbox.I cannot access the django server running in vagrant virtualbox using http://localhost:9991
I have disabled my firewall as well
here's my vagrant file:
VAGRANTFILE_API_VERSION = "2"
def command?(name)
`which #{name}`
$?.success?
end
Vagrant.configure(VAGRANTFILE_API_VERSION) do |config|
# For LXC. VirtualBox hosts use a different box, described below.
config.vm.box = "fgrehm/trusty64-lxc"
# The Zulip development environment runs on 9991 on the guest.
config.vm.network "forwarded_port", guest: 9991, host: 9991, host_ip: "127.0.0.1"
config.vm.synced_folder ".", "/vagrant", disabled: true
config.vm.synced_folder ".", "/srv/zulip"
# Specify LXC provider before VirtualBox provider so it's preferred.
config.vm.provider "lxc" do |lxc|
if command? "lxc-ls"
LXC_VERSION = `lxc-ls --version`.strip unless defined? LXC_VERSION
if LXC_VERSION >= "1.1.0"
# Allow start without AppArmor, otherwise Box will not Start on Ubuntu 14.10
# see https://github.com/fgrehm/vagrant-lxc/issues/333
lxc.customize 'aa_allow_incomplete', 1
end
end
end
config.vm.provider "virtualbox" do |vb, override|
override.vm.box = "ubuntu/trusty64"
# 2GiB seemed reasonable here. The VM OOMs with only 1024MiB.
vb.memory = 2048
end
$provision_script = <<SCRIPT
set -x
set -e
sudo apt-get update
sudo apt-get install -y python-pbs
/usr/bin/python /srv/zulip/provision.py
SCRIPT
config.vm.provision "shell",
# We want provision.py to be run with the permissions of the vagrant user.
privileged: false,
inline: $provision_script
end
How do i access the server from host(Windows)?
I would suggest (on the Guest):
sudo netstat -lnutp
and having a look at what ports are open, and the process which owns them. If the one you want is missing, make sure the service which is responsible for it has been started, or start it yourself. From the looks of your Vagrantfile, this would be the "Zulip development environment".
I am having trouble accessing files through Vagrant on Windows. I have been using it on OS X for quite some time and have my Vagrantfile setup correctly which works every time.
I have sent my colleague the same Vagrant file, he is on Windows and receives 'Permission Denied' when trying to access files through the browser.
Just to be clear, the permission errors are returned by the server when accessing 'dev.local' in the browser and not from Vagrant itself... it will be a configuration error on Windows or within the VM.
The VM is CentOS 6.5
Vagrantfile:
# -*- mode: ruby -*-
# vi: set ft=ruby :
# Vagrantfile API/syntax version. Don't touch unless you know what you're doing!
VAGRANTFILE_API_VERSION = "2"
Vagrant.configure(VAGRANTFILE_API_VERSION) do |config|
config.vm.box = "chef/centos-6.5"
config.vm.network "private_network", ip: "192.168.33.21"
config.vm.network :forwarded_port, guest: 80, host: 8080
config.vm.provision :shell, :path => "install.sh"
config.vm.hostname = "dev.local"
config.vm.synced_folder ".", "/home", id: "vagrant", :nfs => false, :mount_options => ["dmode=777","fmode=777"]
config.ssh.insert_key = false
config.ssh.username = "vagrant"
config.ssh.password = "vagrant"
end
Can any Windows Vagrant users shed any light on this?
It was VBGuestAdditions out of date. The permission error was caused by not being able to sync to my local folder (which contained an index.php) so it was using the servers /home folder which didn't contain anything and since viewing directory structure is disabled it returned permission errors.
I had 4.X.X installed, and VirtualBox is on 5.X.X Here is the fix:
run command: vagrant plugin install vagrant-vbguest
the run vagrant up which may still throw an error as the plugin fails to copy a file.
vagrant ssh to get into the box and run the following command:
sudo ln -s /opt/VBoxGuestAdditions-5.X.X/lib/VBoxGuestAdditions /usr/lib/VBoxGuestAdditions
Replace 5.X.X with the version of your VirtualBox.
logout and run vagrant reload
do a happy dance
Host: Windows 7
Guest: Windows 8
I have a simple Vagrantfile that runs a powershell script to provision the guest. When I packaged the box, I saw that the file was added, but when I run vagrant up I get the error shell provisioner:* `path` for shell provisioner does not exist on the host system: D:/VirtualMachines/test/provision.ps1
I verified that provision.ps1 exists in the vagrant box location under the include directory.
So why isn't provision.ps1 getting copied to the location it needs to when i run vagrant up?
Vagrant file:
VAGRANTFILE_API_VERSION = "2"
modified_name = ENV["COMPUTERNAME"][0..12]
comp_name = modified_name + "TA"
Vagrant.configure(VAGRANTFILE_API_VERSION) do |config|
config.vm.box = "base"
config.vm.hostname = comp_name
config.vm.communicator = "winrm"
config.vm.network "forwarded_port", host: 3389, guest: 3389, auto_correct: true
config.vm.provision "shell", path: "provision.ps1"
end
The answer at How to package files with a Vagrant box? helped me.
Here is how I got it to work:
config.vm.provision "shell" do |s|
p = File.expand_path("../", __FILE__)
s.path = p + "\\provision.ps1"
end
I had the same problem with Vagrant 1.8.1 on host Windows 8. After reading https://github.com/fideloper/Vaprobash/issues/30 I just renamed bootstrap.sh to bootstrp.sh, tried again and it worked. After renaming bootstrp.sh to bootstrap.sh it still worked.
I suppose in my case there was some strange invisible character in the filename.
I have a Vagrantfile as below running under libvirt. When the box boots, the project directory is mounted under "/vagrant" not "/path/to/source". It works fine on another machine under VirtualBox.
Any ideas? Symlinking /vagrant to the actual place I want the mount is hacky.
vagrant up --debug shows that it isn't even attempting it. The port forwarding works OK though.
# -*- mode: ruby -*-
# vi: set ft=ruby :
Vagrant.configure("2") do |config|
config.vm.box = "mybox"
file = File.open("#{Dir.home}/.mybox_key", "rb")
key = file.read
config.vm.network :forwarded_port, guest: 80, host: 8080
config.vm.synced_folder "./", "/path/to/source/", type: "nfs"
end
If you're running an older version of Vagrant that doesn't support this option (e.g. 1.3.5), it may fail silently instead of trying to setup the NFS share. Ensure you're running at least Vagrant 1.5.1.