Which task should be performed first when upgrading host memory?

Using the out-of-band management interface to stop the host would abruptly power it off without migrating VMs, causing workload outages and potential data loss.

Removing the node from the cluster is a destructive operation that permanently removes it from the distributed storage fabric and should never be the first step in a routine hardware upgrade.

Executing 'shutdown -h now' from the AHV CLI immediately powers down the CVM and host without gracefully migrating VMs, causing service disruption.

Placing the node into maintenance mode triggers Nutanix to live-migrate all running VMs off the host to other nodes in the cluster, ensuring workloads remain available and no data is at risk before the node is touched. This step must precede any physical intervention because it coordinates cleanly with AOS and AHV to drain the host in a controlled, non-disruptive manner.