Inventory group variables

Group files

The following files, in the group_vars folder, contain variable definitions for each group of OCP cluster nodes. These group files facilitate more sophisticated settings, such as additional network interfaces.

File	Description
group_vars/bootstrap.yml	Variables defined for the node in the [bootstrap] group
group_vars/infrastructure.yml	Variables defined for all nodes in the [infrastructure] group
group_vars/loadbalancer.yml	Variables defined for all nodes in the [loadbalancer] group
group_vars/nfs.yml	Variables defined for all nodes in the [nfs] group
group_vars/master.yml	Variables defined for all nodes in the [master] group
group_vars/rhcos_worker.yml	Variables defined for all nodes in the [rhcos_worker] group
group_vars/rhel_worker.yml	Variables defined for all nodes in the [rhel_worker] group

Overriding group variables

If you wish to configure individual nodes with different specifications to the ones defined by the group, it is possible to declare the same variables at the node level, overriding the group value. For exanple, if the default CPU and memory resource limits defined in for your worker nodes in the group_vars/worker.yml file are not sufficient for all worker nodes, you can override these values in their respective hosts entries. For example, if these are the CPU, RAM, and disk limits specified at the group level:

cpus: '4'                                               # Number of vCPUs
ram: '16384'                                            # RAM size in MBs
disk1_size: '120'                                       # Disk size in GBs

you can override these values in the individual node entries in the hosts file:

[rhcos_worker]
ocp-worker0   ansible_host=10.15.155.213
ocp-worker1   ansible_host=10.15.155.214
ocp-worker2   ansible_host=10.15.155.215  cpus=8 ram=32768  # Larger worker node for EFK
ocp-worker3   ansible_host=10.15.155.216  cpus=8 ram=32768  # Larger worker node for EFK
ocp-worker4   ansible_host=10.15.155.217  cpus=8 ram=32768  # Larger worker node for EFK

In the example above, three CoreOS worker nodes (ocp-worker2, ocp-worker3, ocp-worker4) are allocated two times more virtual CPU cores and double the RAM compared to the rest of the CoreOS worker nodes.

Common variables across all groups

The following variables apply to all node groups:

Variable	Scope	Description
ip_addr	All nodes	IP address in CIDR format to be given to a node
esxi_host	All nodes	ESXi host where the node will be deployed. If the cluster is configured with DRS, this option will be overridden
cpus	All nodes/groups	Number of virtual CPU cores to assign to a VM or a group of VMs
ram	All nodes/groups	Amount of RAM in MB to assign to a VM or a group of VMs
disk1_size	All nodes/groups	Size of the disk in GB to attach to a VM or a group of VMs.
folder	All nodes/groups	The folder where the VM will be stored
template	All nodes/groups	Override the default template for the node or group. The Bootstrap node ([bootstrap] uses the CoreOS template specified by <<master_template>>. Infrastructure and supporting nodes ([infrastructure], [loadbalancer], [nfs]) use the RHEL template specified by <<infra_template>> by default. The master nodes ([master]) use the CoreOS template specified by <<master_template>>. The CoreOS worker nodes ([rhcos_worker]) use the CoreOS template specified by <<worker_template>>, which is typically the same as <<master_template>>. For RHEL worker nodes ([rhel_worker]), you should set the group default to name of the RHEL template you typically use.
ova_path	All nodes/groups	Name of the OVA used to import the template. The Bootstrap node ([bootstrap] uses the CoreOS OVA specified by <<master_ova_path>>. Infrastructure and supporting nodes ([infrastructure], [loadbalancer], [nfs]) use the RHEL OVA specified by <<infra_ova_path>> by default. The master nodes ([master]) use the CoreOS OVA specified by <<master_ova_path>>. The CoreOS worker nodes ([rhcos_worker]) use the CoreOS OVA specified by <<worker_ova_path>>, which is typically the same as <<master_template>>. For RHEL worker nodes ([rhel_worker]), you should set the group default to name of the RHEL OVA you typically use.