vCloud Suite and vCOPs Options

[ncolyer]

VCOPS Community,

I recently read an article over at VMTurbo: http://www.vmturbo.com/vmware-vcops-changes/

Where they mention the VCOPS versions are trying to be more closely aligned to the vCloud Suite Versions.

Our problem is this.

We use nWorks with SCOM today, with ticketing via a SCOM to CA connector to Unicenter.

We also use vKernel for Capacity Management.

We are trying to rationalize our suite today and we think vKernel still has an easier to read capacity screen/reports, but we like the idea of switching from nWorks to VCOPS Standard.

Will this be included in the vCloud Standard Suite or do you always have to buy it seperately? I understand that the foundation version may be free but that doesn't give us any of the dashboarding/analytics which is a huge appeal for troubleshooting problems. It doesn't seem like we need advanced, so don't see a reason to go to the vCloud Advanced Suite.

Also, does anyone have any links to a "What's new" for Capacity Management section of vCOPs in the next release?

[jddias]

Hi,

  To get SCOM integration you have to be at the Advanced level with 5.6 as you need the Custom UI to install 3rd Party Adapters.  Standard 5.6 is not bundled with vCloud Suite Standard.  Foundation DOES include analytics, you just don't get the main dashboard, which really isn't all necessary given that you only see the Health badges anyway and can do that from the Environment Overview and Operations Details dashboards.

  I want to get a better understanding of your objections to vC Ops versus vKernel.  I went out to the vKernel website and looked at some comparisons between vC Ops and there are some items I don't agree with.  Maybe I can provide some clarification?

[ncolyer]

Thank you for the quick reply.

I'm not looking for SCOM integration at this time. This is purely for VMWare Monitoring.

Our goal would be to replace nWorks running on multiple Windows Servers/SCOM/Universal Connector/Perl Script/CA NSM with

VCOPS > CA NSM via SNMP traps

and use VCOPS for all our dashboarding.

We're still evaluating vKernel vs VCOPs but some of the main items we are missing today are the following:

1. vKernel generates a great report broken down by capacity by Cluster.

2. vKernel allows us to specify the size of a sample VM which we want to use to see how many more VMs we can get in our clusters. The report that comes out to us each week uses this value.

3. Wastefinder etc. seems basically the same on both products.

Am I missing something in points 1 and 2? The other thing also is how VCOPs says based on memory we can get hundreds more VMs on a host/cluster when realistically you might only fit another 20-30, so it's hard to guage how accurate that would be. I spoke with some people at VMware who suggested it is based more on demand and workload as well, and not just consumption, which could help, I just need to trust the numbers first.

[jddias]

vC Ops uses a model VM based on the average VM in your environment to determine your capacity remaining.  You can also model this using "What If..." scenarios in vC Ops.  Like, "What if I add 10 more VMs?" or "What if I double the RAM on my hosts in this cluster?"

I understand your point about the numbers for capacity remaining in vC Ops seeming high.  This is because, by default, vC Ops is set to use a consumption model instead of an allocation model.  You can change the settings in vC Ops so that you allow for buffers for resources if you want to make sure you maintain headroom in your environment.

[ncolyer]

I tried changing my memory buffer to 25% and yet it still seems extremely high.

In one case i have a cluster with 11 Hosts.

Total Capacity (memory): 1725 VMs

Powered On VMS: 323

Remaining Capacity: 1402 VMs

There is just no way i could get another 1402 VMs on these hosts based on memory...

Each host has 128GB of memory and they are all over 75% utilized right now.

[admin]

So the question is to understand what is 1402 VM's size.

My understanding is that we create 2 VM profiles....a Small one and a Large one. Small VM might have a weighting of 1 and Large 2.

So you can deploy 1402 small VM's and 701 Large VM's.

Understand a Small VM can be 1vCPU at 2% utilization and a large VM 2vCPU with 20%...it depends on how this was calculated in vCOPs based on your VM Profiles in your environment.

Note that I stand to be corrected on the way I explained htis...just my simple understanding.