Amazon Linux AMI

Amazon Linux AMIs are set to require tty when sudo'ing, which causes problems when Vagrant tries to do things like setup the rsync folder. In order to get around this you can send aws user_data to populate the cloud init config and change this for the ec2-user when the instance is created:

aws.user_data = "#!/bin/bash\necho 'Defaults:ec2-user !requiretty' > /etc/sudoers.d/999-vagrant-cloud-init-requiretty && chmod 440 /etc/sudoers.d/999-vagrant-cloud-init-requiretty\nyum install -y puppet\n"

The user_data script is executed late in the boot up process. When Vagrant sees SSH is available and attempts a sudo command, this script likely has not executed yet. As a work around, this example Jenkins job script retries the vagrant command:

#!/bin/bash

export AWS_ACCESS_KEY=CHANGEIT
export AWS_SECRET_KEY=CHANGEIT

sudo_unavailable() {
    errorOutput=$WORKSPACE/vagrant.err
    vagrant provision 2>$errorOutput
    exitStatus=$?
    cat $errorOutput
    grep -q 'must have a tty to run sudo' $errorOutput
}

set -x

cd $WORKSPACE/virtual-machine

vagrant destroy --force
vagrant up --provider=aws --no-provision 

count=0
while sudo_unavailable ; do
    sleep 30
    let count++
    if [ $count -gt 4 ]; then
        vagrant destroy --force
        exit 1
    fi
done

if [ $exitStatus = 0 ]; then
    vagrant destroy --force
fi
exit $exitStatus