Skip to main content

Upgrading Cluster Deployments

This page walks through a version upgrade of a cluster deployment, outlining the important considerations.

Context

When clustered, the MCE database is being leveraged by multiple MCE instances together. When upgrading, it is therefore important to ensure all running instances of MCE are on the same version, to ensure compatibility.

Individual nodes in the cluster can be stopped without taking the entire deployment offline, allowing them to be upgraded individually while offline. They should not be started again until all other nodes in the cluster have at least been been stopped, and ideally have been updated, too.

Update Walkthrough

Before updating, it is always recommended to back up the Management Center configuration and the MCE database if possible.

Update nodes

When updating the Management Center version for cluster servers, updated nodes one by one.
Each upgraded node should remain offline until all other nodes have also been updated.

Maximize Uptime

To maximize the uptime of your environment, a MCE cluster server hosting the MindLink front end should remain running until all other cluster servers have been upgraded (but remain offline).

This will temporarily put the system into a one-box configuration hosting MCE and the front end, allowing users to remain active for the maximum amount of time.

Update final server

Once all other cluster nodes have been updated to the desired version, the one-box server should be updated to the desired version, too. This will briefly take the environment offline.

After updating, this server may be started. The environment will return to the temporary the one-box deployment scenario, bringing the environment back online and allowing users to interact with the front end again.

Bring cluster nodes back online

Lastly, each of the offline, upgraded cluster servers can be started one by one to reintroduce the cluster deployment scenario.