NCS Core 6.8

After a VM moves to another host, some reads can initially be served remotely. Nutanix automatically detects remote reads, migrates the VM’s data toward the new host, and makes new writes local, restoring data locality and reducing latency. Shadow clones are intended for commonly shared base disks and do not replace this mechanism.

A completed kickoff and validated Bill of Materials confirm that the deployment scope and required components are correct before travel. A completed cluster questionnaire and completed physical-switch configuration confirm that the required cluster and network parameters are available and the network is ready for deployment.

IP addresses allocated for the cluster deployment must be checked for overlap or conflict with addresses already in use on the customer network, including the network that hosts the DNS servers. Address conflicts can prevent required network communication during deployment and operation.

AHV VM High Availability Guarantee mode reserves resources across the surviving hosts to restart failed VMs. When a cluster has a container configured with replication factor 3, that reservation is designed to tolerate two AHV host failures; consequently, all VMs are guaranteed to restart after two simultaneous host failures.

AHV network visualization discovers the physical topology through LLDP or CDP advertisements from the top-of-rack switches. Switch-port configuration and performance statistics are collected through SNMP polling of those switches; neither dependency is configured individually on every AHV host or CVM.