Upgrade or Downgrade Options
This section provides high-level information about the available upgrade and downgrade options.
Upgrade from a Previous Version
Upgrade from a previous version to Juniper Cloud-Native Router Release 23.1 is supported.
Juniper Cloud-Native Router 23.1 release supports only L2 deployments. L3 deployments are not supported for this release.
To upgrade from an older version to Juniper Cloud-Native Router Release 23.1, perform the following steps:
-
Download the tarball, Juniper_Cloud_Native_Router_23.1.tgz, to the directory of your choice.
-
Expand the file Juniper_Cloud_Native_Router_23.1.tgz.
tar xzvf Juniper_Cloud_Native_Router_23.1.tgz
-
Change directory to Juniper_Cloud_Native_Router_23.1.
cd Juniper_Cloud_Native_Router_23.1
-
Load the JCNR docker images to local docker. The images are available in the Juniper_Cloud_Native_Router_<release-number>/images directory.
docker load -i images/jcnr-images.tar.gz
-
Edit the helmchart/values.yaml file. You must edit the values.yaml file to match the current interface configurations on your currently deployed JCNR instance. For example,
#################################################################### # L2 PARAMS # #################################################################### # fabricInterface: NGDU or tor side interface, expected all types # of traffic; interface_mode is always trunk for this mode fabricInterface: - bond0: interface_mode: trunk vlan-id-list: [100, 200, 300, 700-705] storm-control-profile: rate_limit_pf1 # fabricWorkloadInterface: RU side interfaces, expected traffic is only # management/control traffic; interface mode is always access for this mode fabricWorkloadInterface: - enp59s0f1v0: interface_mode: access vlan-id-list: [700] -
Upgrade the current version of JCNR to release 23.1 by issuing the following command:
helm upgrade jcnr helmchart/.
Upgrading from an older release such as Juniper Cloud-Native Router Release 22.4 results
in having a redundant pod—syslog-ng. Delete the older pod manually after
the upgrade by using the kubectl -n jcnr delete pod
<syslog-ng_pod_name> command.
Verify if Upgrade is Successful
Issue the command kubectl get pods -A. The output of the command shows
all of the pods in the Kubernetes cluster in all namespaces. A successful upgrade shows
that all pods display that they are in the running state. For example,
kubectl get pods -A NAMESPACE NAME READY STATUS RESTARTS AGE contrail-deploy contrail-k8s-deployer-7749c5657d-7qpxw 1/1 Running 0 47m contrail contrail-vrouter-masters-2nqp6 3/3 Running 0 80m jcnr kube-crpd-worker-ds-vd7jn 1/1 Running 0 48m jcnr syslog-ng-c4f7b7455-hxvgl 1/1 Running 0 83m kube-system calico-kube-controllers-5f4fd8666-x2q99 1/1 Running 869 (2m1s ago) 46h kube-system calico-node-28w98 1/1 Running 1 (20h ago) 46h kube-system coredns-54bf8d85c7-868ts 1/1 Running 0 46h kube-system dns-autoscaler-7944dc7978-ws9fn 1/1 Running 1 (20h ago) 46h kube-system kube-apiserver-ix-esx-06 1/1 Running 2 (20h ago) 46h kube-system kube-controller-manager-ix-esx-06 1/1 Running 6 (49m ago) 46h kube-system kube-multus-ds-amd64-jl69w 1/1 Running 1 (20h ago) 46h kube-system kube-proxy-qm5bl 1/1 Running 1 (20h ago) 46h kube-system kube-scheduler-ix-esx-06 1/1 Running 7 (49m ago) 46h kube-system nodelocaldns-bntfp 1/1 Running 1 (20h ago) 46h
Downgrade to an Older Version
As of the 23.1 Release of Juniper Cloud-Native Router, there is no procedure for downgrading to an older version. To change from a current version to an older version, you must uninstall the current version and install an older version.