NetEye Cluster on Microsoft Azure¶

How To install a NetEye cluster on Microsoft Azure¶

Create and manage resources on Azure using Terraform¶

The files required by this guide can be found in the repository at https://github.com/WuerthPhoenix/neteye-azure-installation.

Important

To provision the infrastructure, you must have both the terraform and az (Azure) CLI tools installed on your PC.

Warning

Terraform will create a terraform.tfstate file, which contains the configuration of the resources on Azure and some credentials. It must be considered a SECRET and must not be lost.

The terraform files are kept in the directory /src/terraform.
Follow this configuration guide to setup the terraform variables, afterwards you can follow the first part of the README.md file to deploy the resources on Azure.

Terraform variables configuration¶

Login on Azure witn az login (follow the login procedure on Azure Terraform Provider).
Gather the Azure subscription ID with az account list.
Create a file *.tfvars with the following content (make sure you change the variable values as you see fit):

azure_subscription_id = "<The Azure subscription ID from the previous step>"

resource_group_name  = "neteye_group"
resource_name_prefix = "neteye_terraform"
cluster_size         = 2
vm_size              = "Standard_E4as_v5"
disk_size            = 256

The variables are:

azure_subscription_id: the Azure subscription ID
resource_group_name: the name of the resource group in which the resources will be created.
resource_name_prefix: the prefix for the names of all the resources that will be created, including the VMs.
vm_hostname_template: the template to be used to generate the external hostnames of each VM. It must contain the string %02d where the number of the VM must be written (e.g. neteye%02d.test.it for VM 1 will be neteye01.test.it).
cluster_size: the number of virtual machines to be created.
vm_size: the size to be used when creating the virtual machines. Check the Azure documentation for valid values.
disk_size: the size of the data disk in GB.

Provision the resources¶

To start the provisioning process run the following command:

terraform apply --var-file "<file defined previously>.tfvars"

To get the ne_root password use:

terraform output --raw admin_password

Delete the resources¶

To start the deletion process — which is handy for cleanup after creating a test cluster, for example — run the following command:

terraform destroy --var-file "<file defined previously>.tfvars"

Note

Try not to change the configuration of the created resources manually, if you need to make changes modify the code and open a PR.

To correctly delete the created resources you need to run the destroy command from the same place that ran the apply command (it needs to have the same state saved in terraform.tfstate).

Configure the VMs to create a NetEye cluster¶

Warning

There is only one NIC per VM (thus only one subnet). For this reason you must set the NIC as Trusted:

firewall-cmd --set-default-zone trusted

You can verify by checking the presence of eth0 in the interfaces field after running the following command:

firewall-cmd --zone=trusted --list-all

The /etc/hosts file is already populated with both internal and external IPs.

1. Transform RHEL to NetEye¶

Enable the IPs on repo.wuerth-phoenix.com.

Note

If you are < 4.43 also install network-scripts (dnf install network-scripts)

Warning

Disable SELinux:

sed -i 's/SELINUX=enforcing/SELINUX=permissive/' /etc/selinux/config
setenforce permissive

Run (on all nodes) this script: src/scripts/rhel-to-neteye.sh passing the NetEye version. For example:

rhel-to-neteye.sh 4.43

Warning

Restart the shell to populate all the new environment variables: exec bash

2. Follow NetEye Guide until Fencing¶

Warning

Note that the nodes start from index 00 (and not 01, i.e. neteye00.example.it).

At this point you should have more or less a VM bootstrapped with a NetEye ISO. You can follow the guide at Cluster Nodes - NetEye User Guide.

Caution

Terraform tends to override manual changes to resources if you re-run it. Be aware of this behaviour and ensure any manual steps are documented and reapplied as needed.

Please see:

Warning

When you reach the Cluster Fencing Configuration part please run dnf install fence-agents-azure-arm and follow the steps explained in this Red Hat guide to setup fencing.

Afterwards continue with the steps below.

3. Set the nic value on cluster_ip¶

pcs resource update cluster_ip nic=eth0

4. Edit and setup cluster templates¶

Note

For Non PCS-managed Services you can follow the steps on the guide.

Set the correct volume_group, and 10.1.0 as ip_pre.

Warning

Don’t change the default ip_post value.

Run the Perl script as described in the NetEye Guide.

5. Add azure-lb pcs resources¶

You can run the src/ansible/azure-lb-pcs-resources.yml Ansible playbook (on one node).

Warning

If you run this playbook multiple times, the last two tasks (Add cluster ip res and Add colocation) will fail on subsequent runs because the resources already exist. This is expected behavior.

6. Proceed with regular configuration¶

You can continue following the NetEye Guide as usual from Cluster Nodes - NetEye User Guide onwards.