HA, or, ‘high availability’, mode refers to high availability of the BIG-IP device(s). The F5 Agent for OpenStack Neutron can configure BIG-IP to operate in standalone, pair, or scalen mode. The F5 agent configures LBaaS objects on HA BIG-IP devices in real time.
standalone
mode.Edit the F5 Agent Configuration File
Use your text editor of choice to edit the F5 Agent Configuration File as appropriate for your environment.
vim /etc/neutron/services/f5/f5-openstack-agent.ini
Set the Device driver settings.
Set f5_ha_type
as appropriate for your environment.
standalone
: A single BIG-IP devicepair
: An active-standby pair of BIG-IP devicesscalen
: An active device service cluster of 2 to 4 BIG-IP devices#
# HA mode
#
f5_ha_type = standalone
#
Set up the F5 Agent for OpenStack Neutron to use L2-adjacent mode or Global Routed mode.
High availability modes provide redundancy, helping to ensure service interruptions don’t occur if a device goes down.
standalone mode utilizes a single BIG-IP device; here, ‘high availability’ means that BIG-IP core services are up and running, and VLANs are able to send and receive traffic to and from the device.
pair mode requires two (2) BIG-IP devices and provides active-standby operation. When an event occurs that prevents the ‘active’ BIG-IP device from processing network traffic, the ‘standby’ device immediately begins processing that traffic so users experience no interruption in service. The standby device takes over the entire traffic load, avoiding a loss in performance.
Scalen allows you to configure multiple active devices, each of which can fail over to other available active devices (active-active mode). For example, if two BIG-IP devices are using active-active mode, both devices in the pair actively handling traffic. If an event occurs that prevents one device from processing traffic, that traffic automatically directs to the other active device.
Note
When failover occurs on an active-active cluster, a secondary device takes over the peer traffic load in addition to its current load. Depending on device configuration and capabilities, there may be a reduction in performance.