A running AHV VM hosts a database that developers need to test against. You need a copy of the running VM whose application state is consistent, and you must avoid relying on a raw disk copy. Which action should you take?

Nutanix creates application-consistent recovery points by coordinating with the guest through Nutanix Guest Tools (NGT). When an application-consistent snapshot is requested, NGT triggers Microsoft VSS on Windows or pre-freeze and post-thaw scripts on Linux, so the database flushes its buffers and briefly holds writes while the storage snapshot is taken. The resulting snapshot is redirect-on-write and nearly instant, and cloning it to a new VM gives developers a copy whose application state is consistent, without touching the production VM. A crash-consistent snapshot captures the disks at one moment without informing the guest, which is like pulling the power cord: the database will usually recover, but in-flight transactions and cached writes may be lost or need replay, so it does not meet the requirement. Copying the virtual disks of a running guest is even worse, because the copy is not taken at a single point in time and writes can land halfway through. Live migration only moves the running VM between hosts; it doesn't create a copy at all. Exam caveat: application consistency requires NGT with VSS or quiesce scripts; without NGT the snapshot falls back to crash-consistent. Operational check: confirm NGT is installed and communicating for the VM, then verify the snapshot is marked application-consistent before you clone it.