VMware Cloud Community
txarls_89
Contributor
Contributor
Jump to solution

vRops 6.7 and memory usage

Hi everybody!

We are trying the new version of vmware vRops 6.7 and there are some metric compared with the previous version that are very confuse, one of them is the "memory usage".

In vRops 6.6 the metric "memory usage" is the "guest memory", that's ok. Now you read the release notes about 6.7

"

The Memory|Usage (%) metric of Virtual Machines considers memory usage from Guest OS perspective and not from hypervisor perspective.

In previous releases, the Memory|Usage (%) metric of virtual machines referred to the amount of memory that is actively used, as estimated by VMkernel based

on recently touched memory pages.

This was different from what you would see inside the Guest OS as a memory usage.

The formula of metric Memory|Usage (%) is now changed to (Memory|Utilization (KB) / Memory|Total Capacity (KB)) * 100.  Here the newly introduced Memory|Utilization (KB) metric depends on the Guest OS metric,

which is provided through VMware Tools, and is available since vCenter Server 6.0 Update 1, ESXi 6.0 Update

"

One example comparing vROPSa 6.6 and 6.7 in my enviroment

I have a VM1 with 16384 MB Ram Memory Size

VM1--> vROPS 6.6

Memory usage-->21.79 %

Memory recommendation size--> 7413 MB RAM

Active guest memory-->3768 RAM

VM1-> vROPS 6.7

Memory usage-->87.54 %

Memory recommendation size--> 16777 MB RAM

Active guest memory-->3768 RAM

will this be the topic? Because the are a lot of difference between 7,4 GB and 16 GB in recommendations terms. I have to purchase a new hardware....

Best regards

31 Replies
silus
Enthusiast
Enthusiast
Jump to solution

Thanks for the response on this particular issue.  Will there be any official comment on the otherwise terrible state of 6.7?

As in why was it released in this state?  Why remove so much core functionality? Was this rushed out mid re-development with so much content removed just to hit the vsphere 6.7 release?  Will it be coming back?

6.7 is practically unusable right now.

Reply
0 Kudos
sunnydua2011101
VMware Employee
VMware Employee
Jump to solution

Hi silus

Is it possible to talk about this topic in detail. I will be happy to setup something if you are okay with it. I am available on duas@vmware.com. I will be happy to run you through vROps 6.7 and discuss the issues you have highlighted.

Please reach out to me and we can setup something.

Regards

Sunny

Regards Sunny
Reply
0 Kudos
daphnissov
Immortal
Immortal
Jump to solution

sunnydua201110141​, how about making this a public AMA on vROps 6.7? There are a lot of people unhappy with this release and have many questions. A public forum may be better to clear up any confusion and get some questions answered.

Reply
0 Kudos
silus
Enthusiast
Enthusiast
Jump to solution

I would agree with this, it should be something vmware are making public.

By and large the issues don't affect us now as we have had to revert back to 6.6.1 where we will be stuck until such time as features are put back in that we can make a sideways-step to a new version and not lose functionality.

But it would be useful for everyone to know why it ended up in such a sorry state, and what the plans and timescales are to rectify it.

Considering some baffling schoolboy errors, such as not being able to search/sort/filter the tables in the capacity reclaimable view where that is the one and only place you have put that info and removed it from everywhere else, I can't believe this version even went through any sort of QA before being made public.

Reply
0 Kudos
TacoSauce
Enthusiast
Enthusiast
Jump to solution

There is an official KB now regarding the high memory usage metrics:

VMware Knowledge Base

Reply
0 Kudos
sunnydua2011101
VMware Employee
VMware Employee
Jump to solution

@Silus all the Capacity Reclaimaible views are sortable and searchable. There are additional views and reports available for rightsizing VMs as well. The name of the report is Reclaimation Report or Optimization Report. Please check and revert if you have issues.

Regards Sunny
Reply
0 Kudos
silus
Enthusiast
Enthusiast
Jump to solution

It's not though.  Take for example this workflow that is now broken in 6.7.

A VM has performance issues.  Where can you see how much resource to add/remove?  It used to be on the vm. Now if you want to see, you can go to optimize performance > overview where they are listed in two tables for undersized or oversized VMs  (As an aside, why is the most useful info about capacity reclaimable from vms under the optimize performance view, instead of under capacity reclaimable??) .

The problem is those two tables are not sortable or searchable, at least in the 6.7 instances i've installed. So if you have hundreds of vms requiring attention you have to manually keep clicking through the pages to find the vm you are looking for and what actions to take.

10 steps backwards from being able to see it on the vm in 6.6.

jmartiniber
Enthusiast
Enthusiast
Jump to solution

Any news?

...

-------------------- Please consider marking my answer as "helpful" or "correct"
Reply
0 Kudos
skruje
Contributor
Contributor
Jump to solution

I would love an update on this as well.

Reply
0 Kudos
MeImNot76
Enthusiast
Enthusiast
Jump to solution

Hello,

I had this issue and opened an SR with VMware, they recommended disconnecting/reconnecting all the hosts in the environment (yeah a very big pain)...this worked! When comparing what vROPS is saying as Memory Usage at O/S level it still doesn't compare 100% to what the VM O/S is saying but pretty close, to get even better results you'd need to install the latest VMware tools (version 10.3.0). Obviously I wont be installing the vmware tools on the appliances (NSX, vCenter stuff) so I cannot say that I'm 100% happy but at least the dashboards have calmed down from all that red Smiley Happy

Hope this helps!

Reply
0 Kudos
redsnapper76
Contributor
Contributor
Jump to solution

We are phasing the same issue

We recently upgraded to version 7, where a "new" view was introduced 'business week workload' or Stress in previous versions.

All our vm's are running FULL RED, which confuses our support guys 

As the counter in the workload detail shows demand fully in red, this counter must be coming from "memory host demand" as here it shows also 100%

Capacity = 36 GB
Demand = 36 GB and demand is fully in red

Capacity = 36 GB
Demand = 36 GB

Is there a way to change the "business week workload" view memory counter to something like Guest Memory Active, which is much more acurate to the real situation

Reply
0 Kudos
ooajala
Enthusiast
Enthusiast
Jump to solution

Hi Sunny,

I was able to enable the old metric in policies; I see it now when I select All Metrics on a virtual machine but when I try to edit a View and I select the Data section > The Memory|Guest Active Memory (%) does not show.

pastedImage_1.png pastedImage_2.png

Any ideas?

--
*** If you like my response, please click "Like" below, if you think I answered your question to the best of my abilities, please mark post as Correct Answer ***

3x vROps Environments implemented across my organization. VDI vROPs 8.4 ::: Epic vROPs 8.4 ::: Shared vROPs 7.0
vROPs Adapters in use: EpicCare & Horizon View Adapter 2.1
Reply
0 Kudos