Deploying a HDInsight Cluster with an insufficient head node size will result in a silent head node size increase

[Base source code](https://github.com/Azure/azure-quickstart-templates/tree/master/101-hdinsight-spark-linux)

### Issue Details

As per [HDInsight doc](https://docs.microsoft.com/en-us/azure/hdinsight/hdinsight-hadoop-provision-linux-clusters#virtual-machine-sizes), it says "If you need more than 32 worker nodes in a cluster, you must select a head node size with at least **8 cores** and 14 GB of RAM.".

### Repro steps (*if necessary, delete otherwise*)
1. Change vm type:
Change `headnode.targetInstanceCount` from 2 to 1
https://github.com/Azure/azure-quickstart-templates/blob/master/101-hdinsight-spark-linux/azuredeploy.json#L100
2. Create resource group
`az group create --name "$resource_group_name" --location "East US"`
3. `az group deployment create --resource-group "$resource_group_name" --name "$deploy_name" --template-file azuredeploy.json --parameters @azuredeploy.parameters.json`
4. The deployment will succeed with
```
"computeProfile": {
"roles": [
{
"encryptDataDisks": false,
"hardwareProfile": {
"vmSize": "Standard_D12_v2"
},
"name": "headnode",
"osProfile": {
"linuxOperatingSystemProfile": {
"username": "sshuser"
}
},
"targetInstanceCount": 2 # not 1
},

```
The cluster will be launched with `D12 v2 (**x2**), D12 v2 (x5)`. As there is no prompt or warning message during the process, end users may get caught by surprise.

The affected template is 101-hdinsight-spark-linux/azuredeploy.json, especially the headnode.targetInstanceCount setting near line 100. Reproduce the deployment with that value set to 1 and compare the requested and resulting computeProfile values. Confirm with maintainers whether the intended completion is validation, a warning, or documentation that prevents the silent increase.

Written by the indexing model from the issue text.