How to Upgrade Contrail Networking Through Kubernetes and/or Red Hat OpenShift
Starting in Contrail Networking Release 21.3, you can update Contrail Networking through Kubernetes and/or Red Hat OpenShift.
You can use this procedure to update Contrail Networking deployed by the Tungsten Fabric (TF) Operator.
To update Contrail Networking:
- Update manifests with the new container tag.
export CONTRAIL_CONTAINER_TAG=<new tag> ./tf-operator/contrib/render_manifests.sh
Note:Only CONTRAIL_CONTAINER_TAG must have a new tag. The render manifest must be done with all the same exported environment variables used during the initial deployment.
- Update the tf-operator deployment.
kubectl apply -k ./tf-operator/deploy/kustomize/operator/templates/
- Wait and ensure that the tf-operator deployment is updated.
kubectl –n tf get pods -w | grep tf-operator
- Update Contrail Networking resources.
kubectl apply -k ./tf-operator/deploy/kustomize/contrail/templates/
- Wait and ensure that the Contrail Control plane pods are
updated.
kubectl –n tf get pods -w
- Use the contrail-status tool to check the Contrail Networking
status on all the master nodes.
$ contrail-status Must show that all services are active: == Contrail control == control: active nodemgr: active named: active dns: active == Contrail analytics-alarm == nodemgr: active kafka: active alarm-gen: active == Contrail kubernetes == kube-manager: backup == Contrail database == nodemgr: active query-engine: active cassandra: active == Contrail analytics == nodemgr: active api: active collector: active == Contrail config-database == nodemgr: active zookeeper: active rabbitmq: active cassandra: active == Contrail webui == web: active job: active == Contrail vrouter == nodemgr: active agent: active == Contrail analytics-snmp == snmp-collector: active nodemgr: active topology: active == Contrail config == svc-monitor: backup nodemgr: active device-manager: backup api: active schema: backup
- Upgrade the Contrail vRouter components (one-by-one or
by groups).
Choose a node to upgrade and obtain the vRouter daemon name for the node.
kubectl describe node <node name>
Delete the vRouter pod resource by specifying the name of the pod you want to delete.
kubectl –n tf delete pod <vrouter1-vrouter-daemonset-xxxxx>
Wait until the new daemon set is run by kubernetes on a node.
Use
kubectl get pods < >commad.kubectl get pods -n tf | grep "vrouter1-vrouter-daemonset" vrouter1-vrouter-daemonset-77cnz 3/3 Running 0 51m vrouter1-vrouter-daemonset-7rlvf 3/3 Running 0 87m vrouter1-vrouter-daemonset-jrzfm 3/3 Running 0 82m vrouter1-vrouter-daemonset-jvhmj 3/3 Running 0 85m vrouter1-vrouter-daemonset-v4brl 3/3 Running 0 52m
The status is showing Running for all the vRouter daemon sets. The number of daemon set entries depends on the cluster size (that is number of master nodes and worker nodes).
You can also verify the status of a particular daemon set. Obtain the new vrouter-daemonset from the
kubectl describe node <node name>command. Check the status of that particular daemon set using thekubectl get pods -n tf | grep "vrouter1-vrouter-daemonset-XXX"command.
- Verify the vRouter agent status by using the
contrail-statuscommand on the node.Control/Master nodes
$ contrail-status Must show that all services are active: == Contrail control == control: active nodemgr: active named: active dns: active == Contrail analytics-alarm == nodemgr: active kafka: active alarm-gen: active == Contrail kubernetes == kube-manager: backup == Contrail database == nodemgr: active query-engine: active cassandra: active == Contrail analytics == nodemgr: active api: active collector: active == Contrail config-database == nodemgr: active zookeeper: active rabbitmq: active cassandra: active == Contrail webui == web: active job: active == Contrail vrouter == nodemgr: active agent: active == Contrail analytics-snmp == snmp-collector: active nodemgr: active topology: active == Contrail config == svc-monitor: backup nodemgr: active device-manager: backup api: active schema: backup
Worker Nodes
vrouter kernel module is PRESENT == Contrail vrouter == nodemgr: active agent: active