Metric Memory|Current Overcommit Ratio show bad information.

[CristianSalguei]

Hi everyone!

I am working with a specific metric in vROPS 7.5: "Memory|Current Overcommit Ratio" at cluster level object. I think this metric should show me the relation between the physical memory available in the cluster vs the virtual memory allocated by the virtual machines on it right?

My cluster has 9 nodes with 1.5TB of RAM each one, and admission control is configured to tolerate 1 host failure.

This is my cluster info:

If I sum all VMs assigned memory the total is 9441 GB. So how could be possible this overcommit ratio: 0.946

I have these powered off VMS:

(all of them sum 154 GB of memory from a total of 9441 GB)

And here the information that I think is wrong:

That more than 13,6 GB is being used for calculating the overcommit ration I think. But I can't understand the calculation for it, because if I have 154 GB of powered off VMS allocated memory only, how can be so much difference, from 9 GB approx to 13,6 GB of allocated memory adding the powered off?

Thanks!

[KabirAli82 · Accepted solution]

That metrics works fine on my end.

What does the metric Memory|Total Capacity show?

[CristianSalguei]

That is the correct metric. I saw it, but anyway there is missing information that I don't know where it comes from, regarding the amount of memory of the powered of VMS.

But anyway it is from that metric where the information is coming from.

thanks!