Note
This feature is experimental at this time and it has not been fully production tested yet.
Octavia is an OpenStack project which provides operator-grade Load Balancing (as opposed to the namespace driver) by deploying each individual load balancer to its own virtual machine and leveraging haproxy to perform the load balancing.
Octavia is scalable and has built-in high availability through active-passive.
Octavia needs connectivity between the control plane and the
load balancing VMs. For this purpose a provider network should be
created which bridges the octavia containers (if the control plane is installed
in a container) or hosts with VMs. Refer to the appropriate documentation
and consult the tests in this project. In a general case, neutron networking
can be a simple flat network. However in a complex case, this can be whatever
you need and want. Ensure you adjust the deployment accordingly. An example
entry into openstack_user_config.yml
is shown below:
- network:
container_bridge: "br-lbaas"
container_type: "veth"
container_interface: "eth14"
host_bind_override: "eth14"
ip_from_q: "octavia"
type: "flat"
net_name: "octavia"
group_binds:
- neutron_linuxbridge_agent
- octavia-worker
- octavia-housekeeping
- octavia-health-manager
Make sure to modify the other entries in this file as well.
There are a couple of variables which need to be adjusted if you don’t use
lbaas
for the provider network name and lbaas-mgmt
for the neutron
name. Furthermore, the system tries to infer certain values based on the
inventory which might not always work and hence might need to be explicitly
declared. Review the file defaults/main.yml
for more information.
Octavia can create the required neutron networks itself. Please review the
corresponding settings - especially octavia_management_net_subnet_cidr
needs to be adjusted. Alternatively, they can be created elsewhere and
consumed by Octavia.
Special attention needs to be applied to the --allocation-pool
to not have
ips which overlap with ips assigned to hosts or containers (see the
used_ips
variable in openstack_user_config.yml
)
Note
The system will deploy an iptables firewall if octavia_ip_tables_fw
is set
to True
(the default). This adds additional protection to the control plane
in the rare instance a load balancing vm is compromised. Please review carefully
the rules and adjust them for your installation. Please be aware that logging
of dropped packages is not enabled and you will need to add those rules manually.
Note
The default behavior is to download a test image from the OpenStack artifact storage the Octavia team provides daily. Because this image doesn’t apply operating system security patches in a timely manner it is unsuited for production use.
Some Operating System vendors might provide official amphora builds or an organization might maintain their own artifact storage - for those cases the automatic download can be leveraged, too.
Images using the diskimage-builder
must be built outside of a container.
For this process, use one of the physical hosts within the environment.
Install the necessary packages:
apt-get install qemu uuid-runtime curl kpartx git jq python-pip
Install the necessary pip packages:
pip install argparse Babel>=1.3 dib-utils PyYAML
Clone the necessary repositories
git clone https://github.com/openstack/octavia.git
git clone https://git.openstack.org/openstack/diskimage-builder.git
Run Octavia’s diskimage script
In the octavia/diskimage-create
directory run:
./diskimage-create.sh
Upload the created user images into the Image (glance) Service:
glance image-create --name amphora-x64-haproxy --visibility private --disk-format qcow2 \
--container-format bare --tags octavia-amphora-image </var/lib/octavia/amphora-x64-haproxy.qcow2
Note
Alternatively you can specify the new image in the appropriate settings and rerun the ansible with an appropriate tag.
You can find more information abpout the diskimage script and the process at https://github.com/openstack/octavia/tree/master/diskimage-create
Here is a script to perform all those tasks at once:
#/bin/sh apt-get install qemu uuid-runtime curl kpartx git jq pip -v >/dev/null || {apt-get install python-pip} pip install argparse Babel>=1.3 dib-utils PyYAML pushd /tmp git clone https://github.com/openstack/octavia.git git clone https://git.openstack.org/openstack/diskimage-builder.git pushd octavia/diskimage-create ./diskimage-create.sh mv amphora-x64-haproxy.qcow2 /tmp popd popd #upload image glance image-create --name amphora-x64-haproxy --visibility private --disk-format qcow2 \ --container-format bare --tags octavia-amphora-image </var/lib/octavia/amphora-x64-haproxy.qcow2
Note
If you have trouble installing dib-utils from pipy consider installing it directly from souce ` pip install git+https://github.com/openstack/dib-utils.git`
Note
For production installation make sure that you review this very carefully with your own security requirements and potantially use your own CA to sign the certificates.
The system will automatically generate and use self-signed certificates with different Certificate Authorities for control plane and amphora. Make sure to store a copy in a safe place for potential disaster recovery.
In rare cases it might be beneficial to gain ssh access to the amphora for additional trouble shooting. Follow these steps to enable access.
Create an ssh key
ssh-keygen
Upoad the key into nova as the octavia user:
openstack keypair create --public-key <public key file> octavia_key
Note
To find the octavia user’s username and credentials review the octavia-config file on any octavia container in /etc/octavia.
Configure Octavia accordingly
Add a octavia_ssh_enabled: True
to the user file in
/etc/openstack-deploy
Beginning with the Pike release, Octavia can be deployed in a stand-alone
version thus avoiding the Neutron integration. Currently, the following
configuration should be added to openstack_user_config.yml
:
# Disable Octavia support in Neutron
neutron_lbaas_octavia: False
# Disable LBaaS V2
neutron_lbaasv2: False
# Enable Octavia V2 API/standalone
octavia_v2: True
# Disable Octavia V1 API
octavia_v1: False
Please note that in some settings the LBaaS plugin is directly enabled in the
neutron_plugin_base
so adjust this as necessary.
Please be aware that if you enable only the Octavia endpoint, only Octavia load balancers can be created because the integration with 3rd party load balancer vendors nor with the haproxy namespace driver is available in the Pike release.
Please have a close look at the main.yml
for tunable parameters.
The most important change is to set Octavia into ACTIVE_STANDBY mode
by adding octavia_loadbalancer_topology: ACTIVE_STANDBY
and
octavia_enable_anti_affinity=True
to ensure that the active and passive
amphora are (depending on the anti-affinity filter deployed in nova) on two
different hosts to the user file in /etc/openstack-deploy
To speed up the creation of load balancers or in a SINGLE topolgy
to speed up the failover a spare pool can be used.
The variable octavia_spare_amphora_pool_size
controls
the size of the pool. The system will try
to prebuild this number so using too big a number will
consumes a lot of unnecessary resources.
Except where otherwise noted, this document is licensed under Creative Commons Attribution 3.0 License. See all OpenStack Legal Documents.