Some Terraform modules to simply rollout machines on libvirt, AWS and Azure with a unified configuration.
I created this for my own benefit to easily and fast deploy simple virtual machines do check out things as well as integrate it into my testing pipeline. Therefore the features are limited and continuity between all providers was mandatory.
In short, my requirements are:
- deploy machines on libvirt(KVM), Azure and AWS
- only unified(!) configuration options I really need, like:
- telling where the machine should be created (libvirt hypervisor URI, AWS Regions, etc.),
- naming my deployment,
- defining the (SLE) operating system,
- minimal influence on machine sizing (memory, CPUs, disk size, etc.),
- amount of machines.
- ICMP and SSH are allowed
- root access to the machine by ssh key
- logon user with ssh key
- registration on the SCC (or other registration server)
- (a bit more, but this was not doable on libvirt, AWS and Azure all together, so I plan to use Ansible later on)
I'll do bug fixing of cause and also maybe enhance it a bit, but only to the extend of my needs. There are no plans to add stuff for more complex deployments. Feel free to fork and work on your own!
You don't need to clone this repository. It would be enough to reference the git repo in your terraform plan like shown in the examples. Terraform will download and install the modules on initialization.
If you don't need a provider (meaning, libvirt, Azure or AWS), you can skip the installation part for it. Of cause you can't use the module later on.
Download terraform from https://www.terraform.io/downloads, unpack the zip contains only one binary) and put the binary in /usr/local/bin/.
Don't forget to make it executable.
A terraform -version should work and print the downloaded version.
My modules require a terraform version >= 1.1.0. It does not mean, that they won't work with older ones, but this version was the latest, when I started the project.
The steps are also described here: https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html
Download the latest version, unpack and install it:
# curl "https://awscli.amazonaws.com/awscli-exe-linux-x86_64.zip" -o "awscliv2.zip"
# unzip awscliv2.zip
# ./aws/install
If everything went fine, a aws --version should print the downloaded version.
The steps are also described here: https://docs.microsoft.com/de-de/cli/azure/install-azure-cli
rpm --import https://packages.microsoft.com/keys/microsoft.asc
zypper addrepo --name 'Azure CLI' --check https://packages.microsoft.com/yumrepos/azure-cli azure-cli
zypper install --from azure-cli azure-cli
An older version is available in the SLES/OpenLeap repositories.
If you want to use the latest package from the Microsoft repositories, remove the ones from the SLE repos first with:
zypper rm -y --clean-deps azure-cli, before you install the one from the Microsoft repository. Do not update! Otherwiseazwill not work and terminate with:
/usr/bin/python3.6: No module named azure.cli.__main__; 'azure.cli' is a package and cannot be directly executedIf you're already in the mess, reinstall the old SLE package first and follow the steps above:
zypper install --oldpackage azure-cli-<version>
Just install everything you need to run KVM/quemu-based virtual machines. How to install this, is beyond this guide.
Go to https://us-east-1.console.aws.amazon.com/iam/home?region=us-east-1&skipRegion=true#/security_credentials and create an Access Key/Secret Access Key if not one is already present.
Run aws configure on the command line and provide the key Access Key ID, the Secret Access Key and the default region as minimum. The information are stored in ~/.aws/credentials and ~/.aws/config.
To check if authentication works run: aws ec2 describe-instances to list available instances.
Run az login which opens a web browser session. Choose the account, authenticate and return to the command line where you have executed az login. The browser can be closed.
~ # az login
A web browser has been opened at https://login.microsoftonline.com/organizations/oauth2/v2.0/authorize. Please continue the login in the web browser. If no web browser is available or if the web browser fails to open, use device code flow with `az login --use-device-code`.
...
{
"cloudName": "AzureCloud",
...
"id": "<subscription_id>",
...
}
]
Find the id part in the output and set the account: az account set --subscription "<subscription_id>"
Next, create a Service Principal: az ad sp create-for-rbac --role="Contributor" --scopes="/subscriptions/<subscription_id>
az ad sp create-for-rbac --role="Contributor" --scopes="/subscriptions/<subscription_id>"
The underlying Active Directory Graph API will be replaced by Microsoft Graph API in a future version of Azure CLI. Please carefully review all breaking changes introduced during this migration: https://docs.microsoft.com/cli/azure/microsoft-graph-migration
Creating 'Contributor' role assignment under scope '/subscriptions/501de51d-cd4c-4a11-a0b6-3b2f0b6f1393'
The output includes credentials that you must protect. Be sure that you do not include these credentials in your code or check the credentials into your source control. For more information, see https://aka.ms/azadsp-cli
{
"appId": "<appid>",
"displayName": "...",
"password": "<password>",
"tenant": "<tenant>"
}
Hashicorp's documentation states, that for terraform to work, the returned data must be made available by exported variables:
export ARM_CLIENT_ID="<appid>"
export ARM_CLIENT_SECRET="<password>"
export ARM_SUBSCRIPTION_ID="<subscription_id>>"
export ARM_TENANT_ID="<tenant>"
In my case this was not necessary, but I leave this step in here just in case.
How to set this up, is beyond this guide.
Here a brief example how to setup some test machines on libvirt, AWS and Azure. You can find detailed descriptions about how to use the modules here:
Anyways, you need at least a minor understanding of terraform.
# Our test landscapes with two machines per SLES for SAP 15 release on AWS (A), Azure (B) and libvirt (C).
module "test_landscape_A" {
source = "git::https://github.com/scmschmidt/simple_deployment.git//modules/simple_aws"
location = "eu-west-1"
name = "sschmidt-testlandscape-A"
machines = {
1 = ["t3.nano", "sles4sap_15"],
2 = ["t3.nano", "sles4sap_15.1"],
"3a" = ["t3.nano", "sles4sap_15.1"],
"3b" = ["t3.nano", "sles4sap_15.2"],
4 = ["t3.nano", "sles4sap_15.3"]
}
admin_user_key = "ssh-rsa ..."
subscription_registration_key = "INTERNAL-USE-ONLY-f1..."
enable_root_login = true
}
module "test_landscape_B" {
source = "git::https://github.com/scmschmidt/simple_deployment.git//modules/simple_azure"
location = "westeurope"
name = "sschmidt-testlandscape-B"
machines = {
1 = ["standard_b1", "sles4sap_15"],
2 = ["standard_b1", "sles4sap_15.1"],
"3a" = ["standard_b1", "sles4sap_15.1"],
"3b" = ["standard_b1", "sles4sap_15.2"],
4 = ["standard_b1", "sles4sap_15.3"]
}
admin_user_key = "ssh-rsa ..."
subscription_registration_key = "INTERNAL-USE-ONLY-f1..."
enable_root_login = true
}
module "test_landscape_C" {
source = "git::https://github.com/scmschmidt/simple_deployment.git//modules/simple_libvirt"
name = "sschmidt-testlandscape-C"
machines = {
1 = ["micro", "sles_15"],
2 = ["micro", "sles_15.1"],
"3a" = ["micro", "sles_15.1"],
"3b" = ["micro", "sles_15.2"],
4 = ["micro", "sles_15.3"]
}
admin_user_key = "ssh-rsa ..."
subscription_registration_key = "INTERNAL-USE-ONLY-f1..."
enable_root_login = true
}
# Name, size, image and IP address of the deployed machines.
output "test_machines_A" {
value = [
for name, info in module.test_landscape_A.machine_info :
"${name} : ${info.size}/${info.image} -> ${info.ip_address}"
]
}
output "test_machines_B" {
value = [
for name, info in module.test_landscape_B.machine_info :
"${name} : ${info.size}/${info.image} -> ${info.ip_address}"
]
}
output "test_machines_C" {
value = [
for name, info in module.test_landscape_C.machine_info :
"${name} : ${info.size}/${info.image} -> ${info.ip_address}"
]
}
After terraform init and terraform apply we get at the end the output:
...
Apply complete! Resources: 23 added, 0 changed, 0 destroyed.
Outputs:
test_machines_A = [
"sschmidt-testlandscape-A-1 : t3.nano/sles4sap_15 -> 54.246.144.93",
"sschmidt-testlandscape-A-2 : t3.nano/sles4sap_15.1 -> 34.245.42.110",
"sschmidt-testlandscape-A-3a : t3.nano/sles4sap_15.1 -> 63.35.169.23",
"sschmidt-testlandscape-A-3b : t3.nano/sles4sap_15.2 -> 54.195.64.139",
"sschmidt-testlandscape-A-4 : t3.nano/sles4sap_15.3 -> 34.240.0.107",
]
----
test_machines_B = [
"sschmidt-testlandscape-B-1 : standard_b1/sles4sap_15 -> 40.113.123.170",
"sschmidt-testlandscape-B-2 : standard_b1/sles4sap_15.1 -> 40.113.123.58",
"sschmidt-testlandscape-B-3a : standard_b1/sles4sap_15.1 -> 40.113.126.132",
"sschmidt-testlandscape-B-3b : standard_b1/sles4sap_15.2 -> 40.113.123.133",
"sschmidt-testlandscape-B-4 : standard_b1/sles4sap_15.3 -> 40.113.123.40",
]
----
test_machines_C = [
"sschmidt-testlandscape-C-1 : micro/sles_15 -> 172.31.136.48",
"sschmidt-testlandscape-C-2 : micro/sles_15.1 -> 172.31.157.238",
"sschmidt-testlandscape-C-3a : micro/sles_15.1 -> 172.31.178.175",
"sschmidt-testlandscape-C-3b : micro/sles_15.2 -> 172.31.46.93",
"sschmidt-testlandscape-C-4 : micro/sles_15.3 -> 172.31.115.212",
]
- If a machine was shutdowned a
terraform applywill restart it. The new IP is not known and represented by an empty string. Just runterraform refreshto update the IP address.
- Contrary to the other providers a shutdowned machine will not be detected.
- A shutdowned machine would normally get redeployed, because the IP address is not present anymore.
We ignoring changes to the IP address, therefore shutdowned machines get ignored an represented with an empty IP address. If they have been restarted, an additional
terraform applyorterraform refreshis necessary to update the state. Keep in mind, that the newly started machine will get a new IP address!