Overview

Overview of OpenShift Container Platform workflows

Integrating VSP with the OpenShift Container Platform application workflow allows business applications to be quickly turned up and updated by removing the network lag faced by DevOps teams. VSP supports different workflows with OpenShift Container Platform in order to accommodate scenarios where users can choose ease-of-use or complete control using policy-based automation.

Developer Workflow

This workflow is used in developer environments and requires little input from the developer in setting up the networking. In this workflow, nuage-openshift-monitor is responsible for creating the VSP constructs (Zone, Subnets, etc.) needed to provide appropriate policies and networking for pods created in an OpenShift Container Platform project. When a project is created, a default zone and default subnet for that project are created by nuage-openshift-monitor. When the default subnet created for a given project gets depleted, nuage-openshift-monitor dynamically creates additional subnets.

A separate VSP Zone is created for each OpenShift Container Platform project ensuring isolation amongst the projects.

Operations Workflow

This workflow is used by operations teams rolling out applications.

In this workflow, the network and security policies are first configured on the VSD in accordance with the rules set by the organization to deploy applications. Administrative users can potentially create multiple zones and subnets and map them to the same project using labels. While spinning up the pods, the user can use the Nuage Labels to specify what network a pod needs to attach to and what network policies need to be applied to it. This allows for deployments where inter- and intra-project traffic can be controlled in a fine-grained manner. For example, inter-project communication is enabled on a project by project basis. This may be used to connect projects to common services that are deployed in a shared project.

Installation

The VSP integration with OpenShift Container Platform works for both virtual machines (VMs) and bare metal OpenShift Container Platform installations.

Installation for single master

In the Ansible nodes file, specify the following parameters in order to set up Nuage VSP as the network plug-in:

 # Nuage specific parameters
 openshift_use_openshift_sdn=False
 openshift_use_nuage=True
 os_sdn_network_plugin_name='nuage/vsp-openshift'
 openshift_node_proxy_mode='userspace'
 nuage_openshift_monitor_rest_server_port=9443

 # VSP related parameters
 vsd_api_url=https://192.168.103.200:8443
 vsp_version=v4_0
 enterprise=nuage
 domain=openshift
 vsc_active_ip=192.168.103.201
 vsc_standby_ip=192.168.103.202
 uplink_interface=eth0

 # rpm locations
 nuage_openshift_rpm=http://location_of_rpm_server/openshift/RPMS/x86_64/nuage-openshift-monitor-4.0.X.1830.el7.centos.x86_64.rpm
 vrs_rpm=http://location_of_rpm_server/openshift/RPMS/x86_64/nuage-openvswitch-4.0.X.225.el7.x86_64.rpm
 plugin_rpm=http://location_of_rpm_server/openshift/RPMS/x86_64/vsp-openshift-4.0.X1830.el7.centos.x86_64.rpm

 # Optional parameters
 nuage_interface_mtu=1460
 nuage_master_adminusername=admin
 nuage_master_adminuserpasswd=admin
 nuage_master_cspadminpasswd=admin
 nuage_openshift_monitor_log_dir=/var/log/nuage-openshift-monitor

 # Required for brownfield install (where a {product-title} cluster exists without Nuage as the networking plugin)
 nuage_dockker_bridge=lbr0

Installation for multiple masters (HA)

An environment with High Availability (HA) can be configured with multiple masters and multiple nodes.

Nuage VSP integration in multi-master mode only supports the native HA configuration method described in this section. This can be combined with any load balancing solution, the default being HAProxy. The inventory file contains three master hosts, the nodes, an etcd server, and a host that functions as the HAProxy to balance the master API on all master hosts. The HAProxy host is defined in the [lb] section of the inventory file enabling Ansible to automatically install and configure HAProxy as the load balancing solution.

In the Ansible nodes file, the following parameters need to be specified in order to setup Nuage VSP as the network plug-in:

 # Create and OSEv3 group that contains masters, nodes, load-balancers, and etcd hosts
 masters
 nodes
 etcd
 lb

 # Nuage specific parameters
 openshift_use_openshift_sdn=False
 openshift_use_nuage=True
 os_sdn_network_plugin_name='nuage/vsp-openshift'
 openshift_node_proxy_mode='userspace'

 # VSP related parameters
 vsd_api_url=https://192.168.103.200:8443
 vsp_version=v4_0
 enterprise=nuage
 domain=openshift
 vsc_active_ip=192.168.103.201
 vsc_standby_ip=192.168.103.202
 uplink_interface=eth0

 # rpm locations
 nuage_openshift_rpm=http://location_of_rpm_server/openshift/RPMS/x86_64/nuage-openshift-monitor-4.0.X.1830.el7.centos.x86_64.rpm
 vrs_rpm=http://location_of_rpm_server/openshift/RPMS/x86_64/nuage-openvswitch-4.0.X.225.el7.x86_64.rpm
 plugin_rpm=http://location_of_rpm_server/openshift/RPMS/x86_64/vsp-openshift-4.0.X1830.el7.centos.x86_64.rpm

 # Required for Nuage Monitor REST server and HA
 openshift_master_cluster_method=native
 openshift_master_cluster_hostname=lb.nuageopenshift.com
 openshift_master_cluster_public_hostname=lb.nuageopenshift.com
 nuage_openshift_monitor_rest_server_port=9443

 # Optional parameters
 nuage_interface_mtu=1460
 nuage_master_adminusername='admin's user-name'
 nuage_master_adminuserpasswd='admin's password'
 nuage_master_cspadminpasswd='csp admin password'
 nuage_openshift_monitor_log_dir=/var/log/nuage-openshift-monitor

 # Required for brownfield install (where a {product-title} cluster exists without Nuage as the networking plugin)
 nuage_dockker_bridge=lbr0

 # Specify master hosts
 [masters]
 fqdn_of_master_1
 fqdn_of_master_2
 fqdn_of_master_3

 # Specify load balancer host
 [lb]
 fqdn_of_load_balancer