Azure HPC Cluster AutoGrowShrink Not Triggering

I did setup Azure HPC cluster in Azure with 1 head Node and 2 compute Nodes, which is working fine. I am able to submit job from Excel that gets results back perfectly.

Now, I am trying to auto scale the cluster, I have uploaded the Cert to Azure subscription and updated the Cert store and Registry on head node, as described here.

https://azure.microsoft.com/en-in/documentation/articles/virtual-machines-windows-classic-hpcpack-cluster-node-autogrowshrink/

I have set the AutoGrowShrink property using PS, which seems to setup correctly as well.

PS C:\Program Files\Microsoft HPC Pack 2012\Bin> Get-HpcClusterProperty -AutoGrowShrink

Name                                     Value
----                                     -----
EnableGrowShrink                         True
TasksPerResourceUnit                     1
GrowThreshold                            1
GrowInterval                             3
ShrinkInterval                           5
ShrinkIdleTimes                          3
ExtraNodesGrowRatio                      1
GrowByMin                                True
SoaJobGrowThreshold                      50000
SoaRequestsPerCore                       20000

However, when I submit the job that requires more cores, it gets queued up for ever and new compute nodes are not being added to the HPC cluster (AutoGrowShrink not triggering). Any idea what I may be missing.

Solution

Hpc auto grow shrink cannot auto add compute nodes to cluster, you need add the nodes to your HPC IaaS cluster first, then if you enable auto grow shrink, it can stop the nodes when no job, and start the nodes when new job is coming.