Vagrant

By Hashicorp.

box - manage "boxes"
destroy - shut down the VM then delete its stored image?
gem
halt - shut down the VM
init - prepare a directory with a new Vagrantfile
package - shut down the VM, then convert it to a 'package' which can be turned into a box? (Or something)
provision - run just the provisioning (eg, Chef, Puppet...) stage
reload - modify the VM configuration (eg, reapply Vagrantfile), reboot the VM, reprovision the modified parts of the vagrantfile
resume - un-suspend (ie, unhibernate)
ssh - open an SSH shell connection to the VM
ssh-config
status
suspend - hibernate the VM
up - some or all of: copy a VM image to create a new VM, apply configuration to it, boot it

The variable from outside the provisioning block can be accessed as

"#{PATH_VAGRANT}"

The ENvironment Varialbe from the host can be accessed as. This is the way to pass infos to the vagrantfile from the host system. Into the provisioner it then may be passed in as a vagrant variable

  GIT_USERNAME=ENV['GIT_USERNAME']

  # Use shell script to provision
  config.vm.provision "shell", inline: <<-SHELL
  
    echo "#{PATH_VAGRANT}"
  SHELL

Getting the path to the folder of the vagrantfile e.g. in order to copy files from the relative location to the vm.

  PATH_VAGRANT=File.dirname(__FILE__)

  # Use shell script to provision
  config.vm.provision "shell", inline: <<-SHELL

    # echo the full path
    echo -e "Source:     #{PATH_VAGRANT}\.aws\config"
  SHELL

Copying files is done with an own provisioner

PATH_VAGRANT=File.dirname(__FILE__)


# upload the AWS configs
config.vm.provision "file", source: "#{PATH_VAGRANT}\\.aws\\config", destination: "~/.aws/config"
config.vm.provision "file", source: "#{PATH_VAGRANT}\\.aws\\credentials", destination: "~/.aws/credentials"
config.vm.provision "file", source: "#{PATH_VAGRANT}\\.aws\\answerfile", destination: "~/.aws/answerfile"

The escapes, e.g. echo of quotes - must be escaped double times :

\\"

docker exec tomcat8 bash -c "echo -e '<role rolename=\\"manager-gui\\"/>' >> /usr/local/tomcat/conf/tomcat-users.xml"

Howto create a new box

# list vms
vboxmanage list vms

# make sure its the correct one
vboxmanage showvminfo vagrant_default_1566982938109_10062

# create a new box
vagrant package --base vagrant_default_1566982938109_10062  --output devenv.v190819.box

# add box to the known boxes
vagrant box add devenv.v190819 devenv.v190819.box

# list boxes and see the new box added to vagrant
vagrant box list

Now you can reference the new box

# reference the new box
...
Vagrant.configure("2") do |config|
	config.vm.box = "myjenkins"
...

vboxmanage list vms

"<inaccessible>" {5642f0cb-6042-486f-8ea6-7443cb5815b7}
"Jenkins-Vagrant_default_1507401281630_21314" {8e3fd4ba-cda9-4539-9781-bbf7ad4fe423}

vvboxmanage unregistervm Jenkins-Vagrant_default_1507401281630_21314 --delete
vboxmanage unregistervm Jenkins-Vagrant_default_1507401281630_21314 --delete

# clear the global-status cache, if a machine shows up, which is not running
global-status --prune

The complete list: https://github.com/hashicorp/vagrant/wiki/Available-Vagrant-Plugins

VBox additions installation. To enable folder mounting on any OS.

vagrant plugin install vagrant-vbguest
vagrant vbguest

Winnfsd - helps handling NFS mounting problems on WIndows.

vagrant plugin install vagrant-winnfsd

To copy files from the host to the guest.

vagrant plugin install vagrant-scp

Echo multiple lines

mkdir -p /home/vagrant/.docker/
echo -e "
{
 \"proxies\":
 {
   \"default\":
   {
		 \"httpProxy\": \"http://localhost:8500\",
   }
 }
}	
" > /home/vagrant/.docker/config.json

Those scripts will use the vagrant configuration output, to connect to the vagrant machine via putty.

The normal way would be to install the special plugin, made for that :)

vagrant install plugin vagrant-multi-putty 
vagrant putty

If you want to go the long way, you can write an own script to do so

The putty client supports the full color palette and so has a better UX, than the CMD.

Put those files to that folder on the path, e.g. near the “vagrant.exe”

vagrantssh.ps1
A script to start the putty client and connect to the vagrant machine

$puttypath="D:\Programme\Putty\putty.exe"
$vagrantProjectConfigs = & vagrant.exe ssh-config

# parse the vagrant configs
$port = ( $vagrantProjectConfigs | Select-String -Pattern "Port" -SimpleMatch  ) -replace '^[ ]*Port[ ]+(.+)[ ]*', '$1'
$identityfile = ( $vagrantProjectConfigs | Select-String -Pattern "IdentityFile" -SimpleMatch  ) -replace '^[ ]*IdentityFile[ ]+(.+)[ ]*', '$1'
$hostName = ( $vagrantProjectConfigs | Select-String -Pattern "HostName" -SimpleMatch  ) -replace '^[ ]*HostName[ ]+(.+)[ ]*', '$1'

& $puttypath vagrant@$hostName -pw vagrant -P $port -i "$identityfile"

vagrantssh.bat
A helper, to call the powershell script above.

PowerShell -NoProfile -ExecutionPolicy Bypass -Command "& '%~dp0/vagrantssh.ps1'"

Listed here https://github.com/hashicorp/vagrant/wiki/Available-Vagrant-Plugins

vagrant install plugin vagrant-multi-putty 
vagrant install plugin vagrant-scp
vagrant install plugin vagrant-proxy

you can create a tunnel exposing it on the host machine

First forward the ports on Vagrant to Host. This makes the guest-port of vm 4440 available on host port 4440

    # forwarding ports
    config.vm.network "forwarded_port", guest: 4441, host: 4441 # rundeck

Then establish a tunnel. Establishing a tunnel from port 4440 on guest vm to remote machine i-012412412ASF2.

The important part here - is NOT to omit the ip, setting it to 0.0.0.0, so that the ssh tunnel gets exposed on all network adapters, not only localhost and so can be reached from host

ssh ec2-user@i-012412412ASF2 -i ~/.ssh/ec2key.priv.openssh.ppk -fNTC -L 0.0.0.0:4441:localhost:4440

Now you can reach your rundeck application on host under: localhost:4441

which redirects to guest-vm port: localhost:4441

which redirects, via the ssh-tunnel to remote server i-012412412ASF2: i-012412412ASF2:4440

In this example the session-manager is involved to create the tunnel, this part is optional see https://docs.aws.amazon.com/systems-manager/latest/userguide/session-manager-getting-started-enable-ssh-connections.html

When developing code in an IDE on Windows-Host
I would like to have the possibility of running code in the Guest-VM. In an environment as close to the deployment as possible:
e.g. on Kubernetes, integrated with other containers.

File-system watcher:

Requirement:

On the Guest - the changes of “file-system” of code-project must be recognized
and the code-project rebuilt.

Also changes in the IDE on Windows-Host
must trigger changes on the file-system in “Guest-VM”.

Problem:

When sharing a code-project via Oracle-Virtualbox shared-folders
and files where changed on IDE on Windows-Host
unfortunately, NO changes where triggered on Guest file-system-watcher.

When sharing a code-project via Rsync - the requirement was fulfilled.

An example of setting up the environment

• On Guest. Install Cygwin, to get Rsync on Windows
• On Guest. Add the “code-project-folder” to the same folder, where “Vagrantfile” is located
• Configure Vagrantfile to sync “code-project-folder”
• On Guest. Run automatic sync command in one cmd-shell
• Ssh into Vagrant. Run continuos build in one shell
• Configure the Spring Boot application to hot-deploy changes
• Ssh into Vagrant. Run application like Spring Boot application

winget install -e --id Cygwin.Cygwin

cd "VagrantProjectFolder"
mkdir "spring-boot-code-project-folder"

Sync folder `spring-boot-code-project-folder`. Based on `checksum`, not timestamp.

Vagrantfile

...
	config.vm.synced_folder "./spring-boot-code-project-folder", "//mnt/spring-boot-code-project-folder/", 
		type: "rsync", 
		rsync__auto: true,
		rsync__args: ["--verbose", "--ignore-times", "--checksum", "--rsync-path='rsync'", "--archive", "--delete", "-z"],
		id: "spring"

..

Open a new Windows-Terminal Tab and run the continuous sync in there:

vagrant rsync-auto

See https://docs.gradle.org/current/userguide/command_line_interface.html#sec:continuous_build

Skip the tests.
Dont use the daemon.

vagrant ssh
cd "/mnt/spring-boot-code-project-folder"
./gradlew build -xtest --no-daemon --continuous

The Spring boot applications needs to be advised,
to take up any rebuilt app and hot-deploy it.

Add dependency to “devtools”.

build.gradle

dependencies {

	// makes Spring boot hot swap rebuilt jar to tomcat
	developmentOnly 'org.springframework.boot:spring-boot-devtools'
}

vagrant ssh
cd "/mnt/spring-boot-code-project-folder"
./gradlew bootRun --args='--server.port=8888'

Thats also the reason for disk /mnt/d not being mounted correctly

	# give the home permissions back to vagrant 
	# without that -  /home/vagrant is owned by root. 
	# and also /home/vagrant/.ssh/authorized_keys
	# so that no ssh connection works and 
	# vagrant ssh - failes with "Error: Authentication failure. Retrying..."
	sudo chown -R vagrant:vagrant "/home/vagrant"