Refer to the Exhibit.
An administrator has configured a vSphere 6.x DRS cluster as shown in the Exhibit.
Based on the exhibit, which statement is true?
A.
A virtual machine can be powered on in the Test Resource Pool with a 6 GB Memory Reservation.
B.
A virtual machine can be powered on in the Dev Resource Pool with a 8 GB Memory Reservation.
C.
A virtual machine from both the Test Resource Pool and the Dev Resource Pool can be powered on with a 4 GB Memory Reservation.
D.
No more virtual machines can be powered on due to insufficient resources.
Explanation:
Exam Dumps have answer A as right.
Can somebody explain why?
Used reservation DevTest = 5GB ( Dev 4G + Test 1G )
Remaining reservations DevTest = 3GB (8-5)
We can borrow 4GB from grandparent DRS Cluster (12-8=4).
4+3=7GB max for new VMs.
A= 7GB – NO
C= >7GB – NO (4+4=8)
D= wrong. we have 7 total, 3 in existing reservations.
Details:
We can see:
TestVM2 never powered on, only 1G reserved in TEST and VM1 is on.
DevVM1 and DEVVM2 have been on @ 4G (2+2)
What happened?
A= 7GB – NO
C= >7GB – NO (4+4=8)
D= wrong. we have 7 total, 3 in existing reservations.
———————–
Another way: 12G total – 5 used = 7 Free
You can power on a VM up to 7GB.
OK, the DB must not like my Greater than / LEss than
Used reservation DevTest = 5GB ( Dev 4G + Test 1G )
Remaining reservations DevTest = 3GB (8-5)
We can borrow 4GB from grandparent DRS Cluster (12-8=4).
4+3=7GB max for new VMs.
A = Under 7GB – YES
B = Over 7GB – NO
C = Over 7GB – NO (4+4=8)
D = wrong. we have 7 total, 3 in existing reservations.
Details:
We can see:
TestVM2 never powered on, only 1G reserved in TEST and VM1 is on.
DevVM1 and DEVVM2 have been on @ 4G (2+2)
———————–
Another way: 12G total – 5 used = 7 Free
You can power on a VM up to 7GB.
A = Under 7GB – YES
B = Over 7GB – NO
C = Over 7GB – NO (4+4=8)
D = wrong. we have 7 total, 3 in existing reservations.
Used reservation DevTest = 5GB ( Dev 4G + Test 1G )
**sorry for the multiple posts, cant edit and didn’t expect the site to choke on my > and < chars.
A is correct because all Resource Pools are Expandable up to the total of 12 GB for the Cluster.
6 GB is already reserved (whether powered on or not), leaving 6 GB Free of the 12 GB total.
Option A is trying to reserve 6GB. 6GB new + the 6GB already reserved = 12GB. That works.
Option B is trying to reserve 8GB which totals 14.
Option C is trying to reserve 8GB (4+4) which totals 14.
Option D is wrong, because Option A can be done.
My gut says this (^RC^) is the best explanation:
If the parent resource pool also has the Expandable Reservation option selected, it can borrow resources from its parent resource pool. Borrowing resources occurs recursively from the ancestors of the current resource pool as long as the Expandable Reservation option is selected.
https://pubs.vmware.com/vsphere-60/topic/com.vmware.vsphere.resmgmt.doc/GUID-8D813BB8-CE07-40F2-B2CA-269C1FB39475.html
Also:
https://blogs.vmware.com/performance/2016/11/expandable-reservations-resource-pools.html
CPU example:
https://pubs.vmware.com/vsphere-60/topic/com.vmware.vsphere.resmgmt.doc/GUID-76691829-06AD-408A-98AE-6D58555936F5.html
The question is, as stated below by turbo.dabar test, is what the behavior is on OFF state VMs w reservations. Are they reserved, thus lowering the RP ExMR or not?
how do we have 6 GB reserved? you mentioned its reserved 6GB regardless of any vm on/off, so i am assuming you are talking about the first RP “TestDev”, wouldnt that be 8GB instead of 6GB ?
You add up the amount of memory for each VM 1+1+2+2 = 6GB Reserved.
Memory of powered off VM’s is however not technically reserved. The reservation only kicks in when you try to start the VM:
=====
Failed to power on VM.
Could not power on VM : Admission check failed for memory resource See the VMware ESX Resource Management Guide for information on resource management settings.
Group vm.4238276: Invalid memory allocation parameters for VM vmm0:New_Virtual_Machine. (min: 20971520, max: -1, minLimit: -1, shares: -3, units: pages)
Group vm.4238276: Cannot admit VM: Memory admission check failed. Requested reservation: 21113383 pages
SharedArea: Unable to find ‘testSharedAreaPtr’ in SHARED_PER_VM_VMX area.
=====
In my opinion, this set of questions (there are multiple iterations in the dumps) is wrong, or is at least phrased wrong.
There are 3GB of powered on reservations. All resource pools are expandable, so we have 9GB left to work with. Booting an 8GB VM or 2 4GB VM’s should not pose a problem. I personally would not do it, because the margins are extremely small when you take into account overhead and host resources, but it should be possible.
Kevin,
I disagree with your take that the question is being misremembered. I just took the exam and was baffled by this question (which appears here exactly as I remember it). While allowing for the possibility of human error, I read this question several times trying to figure out what I was missing because three of the answers should work. All of the RP reservations are expandable. A VM in each RP is on while the other is off. The only obvious wrong answer is that there are not enough resources to power on the VM.
Schmoopie,
I strongly agree with you, I also attempted the exam lately and had this exact question (it is shown here exactly as it was on real exam), the only certain wrong answer here is that there are no more resources, as for the rest any of them is correct in my opinion.
I concur that technically all three – A, B and C – are correct.
In all likelihood, there’s an error in the provided exhibit. For example, if Dev RP is non-expandable, then the only correct answer is ‘A’ indeed !
I’m gong to contact VMware Education to see what they have to say (most likely their typical “we don’t comment on our exams’ content”…
Tested in a demo environment and correct answer is A.
What I did not know is that the Expandable memory takes the reservation each time a VM is started. Even after you turn it off, the slider in resource pool configuration, where you set the Expandable Memory size cannot go bellow the number of the resources needed from the previous state. Therefore, in Test, one is powered on and the other, evidently, was never powered on. If it was, then the Expandable memory reservation would move to 2GB. On the other hand and here is the catch, the Production pool, one is on and one is off, as shown in the figure (and what is not shown is that sometime it was powered on and that is why the expandable memory rose to 4GB).
So, the only place to add machine even if you turn everything on and the reservation expands as I explained above, where we have total capacity of 12GB in the root, it is a VM from 6GB in Test Pool. You will then have 1+1+6=8GB and 2+2=4GB in Prod Pool. Totals up to 12GB. That is why you can add a VM in Test Pool from 6GB and not from 8GB, nor the option C where you will have 1+1+4=6GB and 2+2+4=8GB which is total of 14GB.
Puzzle solved ! First discovered this while bending my head around Q170 in this dump. Turns out reservations work differently on the VM level and RP level. VM’s reservations are not active as long as VM is powered off. In contrast, RP’s reservations become active –immediately–, even though there’s no running VMs or – get this – there’re no VMs at all ! For more details and respective backing info see Q170.
So, in this scenario, two RPs have 1GB (Test RP) + 4GB (Dev) = 5GB reserved, thus leaving 12 – 5 = 7GB unreserved RAM. Which evidently means ‘A’ is the only correct answer.
I believe that A and B would work, going from what Andy says on this question and described in Q170. If you put a VM in the Test RP it would be 12-5(1T +4D), which would leave 7GB to play with or answer A. However, according to your description, if you put a VM in the Dev RP is would be 12-3(1T +2D(only 2 of the 4 reserved actually used if you are within that RP), which would leave 9 to play with or answer B. However, I am basing my answer on my interpretation of what you said in your posts. I may be misunderstanding something. Either way, I think the question could be done better. Question 170 clearly follows your explanation, but this one as depicted does not IMHO. Any thoughts?
SO A, B, and C actually!
Yes JM, – agree with you and admit that the explanation that works for Q170 does not apply here. It would work only if additional VMs with specs in B and C were attempted to power on outside of these two pools.
What a bad question.
Reservations only take effect for vms when they are powered on: http://www.yellow-bricks.com/2010/03/03/cpumem-reservation-behaviour/ We only have 2 powered on here.
When applying this to a DRS cluster, DRS only looks at the configured memory, not the reservations: http://frankdenneman.nl/2009/12/08/impact-of-memory-reservation/ We are not told how much memory is configured for each vm.
HA looks at configured reservation, but is not mentioned in the question.
I would choose A, only out of caution. Maybe there is something related to slot size or something else I haven’t found.
Does one have to assume contention? If so then D is the only right answer.
What a crappy question.
Tested this again and again and again in a lab…
my test host has 64 GB so first I created 12 GB reservation RP (non-exp.) with a 12 GB limit, which acts as a DRS cluster in the exibit.
under that, created 8 GB reservation RP (exp., no limit)
under that, created two RP-s, 4 GB + 1 GB, both exp + no limit
finally created 2 VMs in each of child pools (4 total), and set their memory size and memory reservations as shown in exibit, and powered on one 1gb reserved and one 2gb reserved machine.
after that I went to eliminate “B” as an answer so i created new, fifth machine with 8 GB of RAM and with 8 GB reservation, and put it in “DEV” pool. It powered on just fine. After that i created a few more VM-s in every pool, to test how much memory can be allocated, and in every combination, it was up to 12 GB of reserved total memory, no matter what the RP memory reservations was as long as it was expandeable and without limit set.
in conclusion, I think this is a poorly formulated question to which A+B+C are all correct answers because I cannot find any reason why we wouldn’t be able to power on machines while we are under the 12 GB total memory as long as all the RPs are expandeable.
to clarify further, it was up to 12 GB of reserved total memory FOR POWERED ON VMs. power off VMs are not counted, and as long as RPs are with expandeable reservations we are fine up to 12 GB of total powered on VMs and reservations on them
I think that falls in line with this post:
To put it simply, a resource pool with ‘expandable reservation’ can expand its reservation by asking more resources from its parent .
The need to expand reservation comes from the increase in reservation demand of its child objects (VMs or resource pools). If the parent resource pool is short of resources, then the parent expands it reservation asking resources from the grand parent.
https://blogs.vmware.com/performance/2016/11/expandable-reservations-resource-pools.html
In case of failover, VM is being restarted at root RP and migrated later to the appropriate RP. But even without HA. How many hosts we have in this cluster? 2? 3? 6?
If 2x6GB or 3x4GB, can we use in this answer all the host memory for VMs? Confusing. Anyway with 2x6GB cluster, max VM size is 6GB. If not all memory is available for VM-s – no 8, no 6, no 2×4 (one 2GB is already running). If all memory – both 6 and 2×4. Or maybe it is a one node DRS cluster – in my lab I have 2 ESXi with 96GB each and DRS cluster memory is 192GB, but some memory from this value is used by ESXi.
I have no idea what you’re talking about… so here’s a bunny with a pancake on its head
http://i.imgur.com/mLaAr7G.jpg