Incorrect error message while creating instance snapshot without memory for a stopped VM

### problem

Incorrect error message while creating instance snapshot without memory for a stopped VM (ROOT & DATA disk on different PowerFlex pools)

```
2026-04-10 09:43:05,535 DEBUG [c.c.v.s.VMSnapshotManagerImpl] (qtp698741991-25:[ctx-8948ca15, ctx-1e54ecb4]) (logid:ea105dce) Creating a snapshot of a running KVM Instance without memory is not supported
2026-04-10 09:43:05,536 INFO [c.c.a.ApiServer] (qtp698741991-25:[ctx-8948ca15, ctx-1e54ecb4]) (logid:ea105dce) Failed to create Instance Snapshot: Creating a snapshot of a running KVM Instance without memory is not supported
```

### versions

ACS 4.20.3 RC4 + Advanced Zone + KVM hypervisor

Should exist with 4.20.1 & 4.20.2 as well, as per the changes here:
https://github.com/apache/cloudstack/pull/9773/changes#diff-7d772616388f0b97219416327c6db1bcad3c162d0271f245ee903198913aa0d3

### The steps to reproduce the bug

1. Deploy a VM with data disk (ROOT & DATA disk on different PowerFlex pools)
2. Stop the VM
3. Take VM snapshot

### What to do about it?

Expecting correct state of the VM in the error message.

Instance snapshots without memory is supported with PowerFlex volumes when all volumes are on the PowerFlex storage and different pools (this seems broken here: https://github.com/apache/cloudstack/pull/3724/changes#diff-705d7f4f3bd1b9d599a38327738fab759b6f6a925f405cbbb7de7bef9b586be2) - now, it is supported when all the volumes are on the same PowerFlex pool only.

Start in VMSnapshotManagerImpl, where the log reports the incorrect running-VM message, and compare the behavior with the changes referenced in PRs 9773 and 3724. Reproduce the stopped-VM snapshot case with ROOT and DATA disks on different PowerFlex pools; done means the error reports the VM's actual state and the supported operation behaves correctly.

Written by the indexing model from the issue text.