Inventories

1. Inventories#

1.1. Create first inventory#

You can now create your first cluster inventory, which acts as a text/folders based database of your cluster description.

To do so, you can either create it manually from scratch using the remaining documentation, or initialize it with the bluebanquise-manager tool. Both methods are exposed bellow.

1.1.1. Create inventory manually#

Create inventories folder, and then create inside this path a new folder with your desired inventory name. In this example, we will call our inventory default.

mkdir -p /var/lib/bluebanquise/inventories/
mkdir /var/lib/bluebanquise/inventories/default/

Now create the group_vars all Ansible structure inside the inventory, along with the cluster folder.

mkdir -p /var/lib/bluebanquise/inventories/default/group_vars/all/
mkdir -p /var/lib/bluebanquise/inventories/default/cluster

Now, set your cluster domain name, and the cluster timezone (tune according to your needs).

Note

You can get the full list of available time zones using command timedatectl list-timezones.

echo 'bb_domaine_name: "bluebanquise.cluster.local"' > /var/lib/bluebanquise/inventories/default/group_vars/all/dns.yml
echo 'bb_time_zone: "Europe/Brussels"' > /var/lib/bluebanquise/inventories/default/group_vars/all/time.yml

Now create your first network. Current configuration is basic, you will be able to add more elements later. Our first network will be called net-admin. Please note that the prefix net- is mandatory here. This will be explained in the networks section of this documentation.

Note also that we will consider here that our primary management node will be on ip 10.10.0.1.

cat << EOF > /var/lib/bluebanquise/inventories/default/group_vars/all/networks.yml
networks:
  net-admin:
    subnet: 10.10.0.0
    prefix: 16
    services_ip: 10.10.0.1
EOF

Now create your nodes file, with our first management node. We will consider that it’s network interface connected on the net-admin network is named enp0s8.

cat << EOF > /var/lib/bluebanquise/inventories/default/cluster/nodes.yml
all:
  hosts:

    # Management nodes
    mgt1:
      # bmc:
      #   name: bmgt1
      #   ip4: 10.10.100.1
      #   network: net-admin
      #   mac: 2a:2b:3c:4d:5e:6f
      network_interfaces:
        - interface: enp0s8
          ip4: 10.10.0.1
          network: net-admin
          mac: 1a:2b:3c:4d:1e:9f
EOF

Finally, create groups files, to register our management node in 3 groups: 1 function group (fn_), 1 os group (os_), and 1 hardware group (hw_). Customize groups names to your needs, but make sure fn_management is preserved. This is a specific naming to identify management servers, and your management nodes should always be in this group.

cat << EOF > /var/lib/bluebanquise/inventories/default/cluster/fn
[fn_management]
mgt1
EOF

cat << EOF > /var/lib/bluebanquise/inventories/default/cluster/os
[fn_management]
mgt1
EOF

cat << EOF > /var/lib/bluebanquise/inventories/default/cluster/hw
[fn_management]
mgt1
EOF

You can now check that your syntax is valid using the ansible-inventory command:

ansible-inventory -i /var/lib/bluebanquise/inventories/default/ --graph

If you don’t see any errors, then your first inventory is ready. You can now customize it using next parts of the documentation.

1.2. Where to set variables#

BlueBanquise stack relies on Ansible to push configuration. As so, it makes usage of the Ansible inventory structure and its precedence mechanism.

The inventory tree is simple:

inventory_root/
├── cluster
├── group_vars
│   └── all
└── host_vars

You can create yaml file anywhere in the inventory, Ansible will read them at execution. (So don’t do backups like myfile.yml.bkp in the inventory, it will be read too!)

Now, see the concept as the following: when you push configuration on a node, the node will inherit from the inventory. BUT, depending of the node, inventory inherited might be different. See that as different persons looking at an object, but from different angles. They see something similar, but not exactly the same.

Explanations:

  1. A node defined in the inventory is by default member of group “all”. So it will inherit of all files stored into group_vars/all/ folder.

  2. A node can be member of a group (or multiple groups). It will then inherit of all the files stored in the group(s) folder(s). For example, if the node is part of group “red”, it will inherit of all files stored into group_vars/red/ folder.

  3. A node can inherit from its dedicated files. If the node name is “c001”, then it will inherit of all files stored into host_vars/c001/ folder.

  4. A node can also inherit from variables specified under its name into cluster/nodes.yml file.

Now that a node has inherited files, precedence mechanism takes place:

  1. If a variable with the same name is defined at 2 places, precedence will tell the node which one to use. Order of priority is the following:

inventory_root/
├── cluster (1)
├── group_vars
│   ├── all (3)
│   └── red (2)
└── host_vars
    └── c001 (1)

So cluster or host_vars (considered same level) win against group_vars/red/ which wins against group_vars/all/.

host_vars > group_vars (!= all) > group_vars/all/

You can use all of that to create and configure your desired cluster, stack will adapt to this. Only limitation is about os_ and hw_ prefixed variables that should only be set in group_vars/X or group_vars/all/ folders (never at host_vars).

Contents