Multicluster engine operator installs and manages multicluster components, such as cluster-manager, hive, and other core foundation resources.
You must meet the following requirements to install the MultiCluster Engine:
- An OpenShift Container Platform (OCP) 4.3+ cluster available
- You must have a default storage class defined
oc
(ver. 4.3+) &kubectl
(ver. 1.16+) configured to connect to your OCP clusteroc
is connected with adequate permissions to create new namespaces in your OCP cluster.
To install the MultiCluster Engine, call the helper script which will supervise the installation of multicluster components. This script requires a SNAPSHOT tag as input.
-
Follow these presteps if deploying MultiCluster Engine downstream builds:
- Follow steps here to set up ImageContentSourcePolicy and configure the global pull secret.
- Set the following environment variable -
export DOWNSTREAM=true
-
Run the
./multiclusterengine/start.sh
script in the root directory of this repository
$ ./multiclusterengine/start.sh
- When prompted for the SNAPSHOT tag, either press
Enter
to use the previous tag, or provide a new SNAPSHOT tag.- UPSTREAM snapshot tags - https://quay.io/repository/stolostron/cmb-custom-registry?tab=tags
- DOWNSTREAM snapshot tags - https://quay.io/repository/acm-d/mce-custom-registry?tag=latest&tab=tags
After the tag has been provided, the installation will continue. Currently the installation deploys and manages its components in the multicluster-engine
namespace which it creates.
You can upgrade you UPSTREAM Multicluster Engine with the upgade.sh script.
- Run the script
# Use the optional --watch parameter ./upgrade.sh
- You will be prompted to enter the snapshot, make sure you choose a
2.0.0-BACKPLANE-...
snapshot - If you added the
--watch
parameter, a watch on the pods will start, so you can see the changes.- You will ALWAYS see the
multicluster-engine-operator-...
pods (two of them) restart. - Any other images that are new to the snapshot will be restarted
- The pods are sorted with NEWEST last
- When the
multicluster-engine-operator
restarts, it will apply any CRD changes
- You will ALWAYS see the
- Press ctrl-c to end the watch!
- Run the script, including the snapshot
export snapshot=2.0.0-BACKPLANE-2022-04-29-01-53-52 && ./upgrade.sh
- The upgrade is performed, but the script does not WAIT for any pod change to occur
To uninstall the MultiCluster Engine, follow these steps:
- Delete the multiclusterengine custom resource -
$ oc delete multiclusterengine --all
- Delete the MultiCluster Engine CSV, Subscription, and namespace.
$ oc delete csv $(oc get sub multicluster-engine -o jsonpath='{.status.currentCSV}')
$ oc delete sub multicluster-engine
$ oc delete namespace multicluster-engine