Edit the file /etc/kolla/globals.yml to add these settings which are specific to kolla-kubernetes:
# Kolla-kubernetes custom configuration
orchestration_engine: "KUBERNETES"
api_interface_address: "0.0.0.0"
memcached_servers: "memcached"
keystone_database_address: "mariadb"
keystone_admin_url: "http://keystone-admin:35357/v3"
keystone_internal_url: "http://keystone-public:5000/v3"
keystone_public_url: "http://keystone-public:5000/v3"
glance_registry_host: "glance"
Then, generate the Kolla configuration files:
# Generate Kolla Configuration Files
pushd kolla
sudo ./tools/generate_passwords.py
sudo ./tools/kolla-ansible genconfig
popd
If using a virt setup, set nova to use qemu unless your environment has nested virt capabilities enabled:
crudini --set /etc/kolla/nova-compute/nova.conf libvirt virt_type qemu
Your cluster needs to have at least one node labeled with each of the following labels:
kolla_compute=true
kolla_controller=true
Label you current node:
ALLINONENODE=$(hostname)
kubectl label node $ALLINONENODE kolla_compute=true
kubectl label node $ALLINONENODE kolla_controller=true
Alternately, you can override the default labeling used in the kolla-kubernetes.yml file. It is also possible to target specific services to specific pools of nodes with this mechanism.
Create the kubernetes namespace. By default it is kolla.
kubectl create namespace 'kolla'
When the namespace is created, each kubectl command executed against the namespace requires adding --namespace=kolla keyword. The following sequence of commands allow setting up the default kubectl context with the right namespace and URL to kube-apiserver thus minimizing amount of typing one needs to do.
kubectl config set-context kolla --namespace=kolla
# X.X.X.X ip address of kubernetes api server
kubectl config set-cluster kolla --server=http://X.X.X.X:8080
kubectl config set-context kolla --cluster=kolla
kubectl config use-context kolla
Before using this script, you MUST generate passwords by using generate_passwords.py (comes with kolla distribution), if there is no password.yml at /etc/kolla, the script will generate an error. Script accepts 1 parameter: create or delete.
# To create Secrets for all services in passwords.yml run:
secret-generator.py create
# To delete Secrets for all services in passwords.yml run:
secret-generator.py delete
Kubernetes uses service discovery for all pods including the net=host pods. In the net=host pods, resolv.conf doesn’t point at kube-dns. Kolla-kubernetes provides a workaround by creating a configmap called resolv-conf with a resolv.conf from a non net=host pod so that dns properly resolves.
Create the resolv.conf configmap:
./tools/setup-resolv-conf.sh
The following commands will walk through the deployment of the OpenStack services. There will be pauses in between commands to sure they completed. In the future, this will be handled as a workflow:
for x in mariadb keystone horizon rabbitmq memcached nova-api \
nova-conductor nova-scheduler glance-api-haproxy \
glance-registry-haproxy glance-api glance-registry \
neutron-server neutron-dhcp-agent neutron-l3-agent \
neutron-metadata-agent neutron-openvswitch-agent \
openvswitch-db-server openvswitch-vswitchd nova-libvirt \
nova-compute nova-consoleauth nova-novncproxy \
nova-novncproxy-haproxy neutron-server-haproxy \
nova-api-haproxy cinder-api cinder-api-haproxy \
cinder-backup cinder-scheduler cinder-volume \
tgtd iscsid; \
do
kolla-kubernetes resource create configmap $x
done
for x in mariadb rabbitmq glance; do
kolla-kubernetes resource create pv $x
kolla-kubernetes resource create pvc $x
done
for x in mariadb memcached keystone-admin keystone-public rabbitmq \
rabbitmq-management nova-api glance-api glance-registry \
neutron-server nova-metadata nova-novncproxy horizon \
cinder-api; \
do
kolla-kubernetes resource create svc $x
done
for x in mariadb-bootstrap rabbitmq-bootstrap; do
kolla-kubernetes resource create bootstrap $x
done
watch kubectl get jobs --namespace kolla
wait for it....
for x in mariadb-bootstrap rabbitmq-bootstrap; do
kolla-kubernetes resource delete bootstrap $x
done
for x in mariadb memcached rabbitmq; do
kolla-kubernetes resource create pod $x
done
watch kubectl get pods --namespace kolla
wait for it...
for x in keystone-create-db keystone-endpoints keystone-manage-db; do
kolla-kubernetes resource create bootstrap $x
done
watch kubectl get jobs --namespace kolla
wait for it...
for x in keystone-create-db keystone-endpoints keystone-manage-db; do
kolla-kubernetes resource delete bootstrap $x
done
kolla-kubernetes resource create pod keystone
watch kolla-kubernetes resource status pod keystone
wait for it...
for x in glance-create-db glance-endpoints glance-manage-db \
nova-create-api-db nova-create-endpoints nova-create-db \
neutron-create-db neutron-endpoints neutron-manage-db \
cinder-create-db cinder-create-endpoints cinder-manage-db; \
do
kolla-kubernetes resource create bootstrap $x
done
watch kubectl get jobs --namespace=kolla
wait for it...
for x in glance-create-db glance-endpoints glance-manage-db \
nova-create-api-db nova-create-endpoints nova-create-db \
neutron-create-db neutron-endpoints neutron-manage-db \
cinder-create-db cinder-create-endpoints cinder-manage-db; \
do
kolla-kubernetes resource delete bootstrap $x
done
for x in nova-api nova-conductor nova-scheduler glance-api \
glance-registry neutron-server horizon nova-consoleauth \
nova-novncproxy cinder-api cinder-scheduler; \
do
kolla-kubernetes resource create pod $x
done
watch kubectl get pods --namespace=kolla
wait for it...
for x in openvswitch-ovsdb-network openvswitch-vswitchd-network \
neutron-openvswitch-agent-network neutron-dhcp-agent \
neutron-metadata-agent-network neutron-l3-agent-network; \
do
kolla-kubernetes resource create pod $x
done
kolla-kubernetes resource create pod nova-libvirt
kolla-kubernetes resource create pod nova-compute
watch kubectl get pods --namespace=kolla
wait for it...
Services should be up now.
If you want to simply access the web gui, see section Web Access below.
This will be automated by an created an “operator pod” in the future. Credentials can be generated by hand by looking in /etc/kolla/globals.yml and filling in these variables:
export OS_PROJECT_DOMAIN_ID=default
export OS_USER_DOMAIN_ID=default
export OS_PROJECT_NAME=admin
export OS_USERNAME=admin
export OS_PASSWORD=<keystone_admin_password>
export OS_AUTH_URL=http://<kolla_internal_fqdn>:<keystone_admin_port>
export OS_IDENTITY_API_VERSION=3
If you want to access the horizon website, fetch the admin password from within the toolbox like:
grep keystone_admin /etc/kolla/passwords.yml
Note
petsets currently arn’t deleted on delete. The resources for it will
have to be cleaned up by hand.