Exam 3V0-21.21 topic 1 question 71 discussion

Definitely B & E.

B & D It is a problem about scare-up scare-out Target: choose the profile ! not design the cluster size or another A:WRONG C : nothing about E:the number of vms in "cluster" ,does there anything different size use two profile ? B : RIGHT or course D: scare-up more power ,cooling ,more license !!! is in design course PPT

Answer B&D A : The manufacturer won't change anything in our case B : Having Profile 1 servers will allow a smaller fault domain in case of host failure, but it may become a limitation if using big VMs. In addition, if we lookup to have only 1 host for failover, the Profile 2 will be more expansive. It has to be considered to choose the hardware. C : Choosing one or the other profile won't impact the downtime allowed D : If we take only the purchase in account, there is no difference between both profiles. But we have also the maintenance to take into account : it may be more expensive with more hosts E : It could be to consider, but we have to consider also the specifications of the VMs (big or small). So the answer is incomplete.

A and e. Everyone is overlooking the fact that some hosts might be mad and others intel and that could severely impact clusters

B: Is clearly the answer E: Number of VMs and host size dictates how easy would it be to migrate VMs from one. host to another. For example, larger VMs are hard to migrate if host is smaller profile.

B,D. explanation, if you take a look to vsphere design the number of VM (E) is not correct, becauase it's not specified the amount of vCPU and vRAM used. You could have 6 huge VM or 200 small VMs; all VM has to share same HW sizing. So, we have: - profile host 1: 10K $ (1 soket 16 cores or 2 socket with 8 cores) - profile host 2: 20k $ (1 socket 32 cores or 2 socket with 16 cores) cost of vcenter standard 5K $ cost vsphere Sandard plus 3K x socket $ cost of Essential Plus: 5K for three hosts dual socket profile 1: you need at least 6 hosts:: 10K x 6 + vCenter 5K + Standard 3K x 6= 83K profile 2: only 3 hosts:: = 20K x 3 + Essential Plus 3K x 1= 63K profile 2 wins in cost to procure and maintain hardware only three hosts that could be licensed with Essential Plus kit

I first thought B and D. But after writing following explanation I am picking B and E, because: A is not relevant for this as we assume architect would pick only from HCL. Also there's no mention of any other requirements which would lead to this one. B is important for determining the resiliency. With profile 1 we have 6 hosts where with profile 2 only 3, so failure domain would be bigger. C is not relevant at all D procurement cost is equal, for both profiles and not sure wat they mean with maintenance (electricity?cooling?operational cost?) but it would be similar as well for both E makes perfect sense with B, also scale-out,scale-up strategy, and vmware mentions this in 7.1 section of Designing vSphere Host Clusters book. Same book also mentions hardware cost, but in combination with spare capacity so in my opinion we have same costs so this is the right choice.

what if the Manufacture and the CPU on host is not supported in HCL in any of the profile ? Should i not be considering A over E - (E is just the number of VMs, for me to decide on the hardware profile how does it matter)

Answer D&E Don't get hung up on cpu model and vendor because it already stated that x2 is expensive VCAP Design document says; The capital costs of purchasing fewer, larger hosts compared to purchasing more hosts that are smaller. Costs vary between vendors and models. • The operational costs of managing a few hosts compared to several hosts. • The maximums for clusters to ensure that you are within the appropriate limits. • The purpose of the cluster, for example: — A virtualized server cluster typically has more hosts with fewer VMs per host. — A VMware Horizon cluster typically has fewer hosts with more VMs per host.

I think is B and D because; they give you the requirements "96 CPU cores and 1.5 TB RAM" So, You need to comply with: The amount of capacity available for failover of virtual machines within the cluster. And, The question talks about HW and costs and with profile 2, you reduce the cost of electricity use, support renewal, etc.

In my opinion A & B is correct. A) You have not being provided with CPU details of each profile, so could be that Profile 1 is AMD and profile 2 is Intel and they will not work together in the same cluster. Even EVC could fix "same manufacturer different model" problem but has to be took in consideration. B) If you configure a "3 nodes cluster with Profile 2 (96CPUs, 1.5TB RAM)" and have a failover event, you will impact 33% of your platform. If you configure a "6 nodes cluster with Profile 1 (96CPUs, 1.5TB RAM)" a single failover event will only impact around 16% of your platform C) for me plays no role while selecting hardware profile. D) is not a problem as Profile 2 is twice the price but also twice of the resources. E) I think is important but not as important as A and B as nobody told us the CPU and RAM ratio of the VMs, so it could be that it will run just fine with 10:1 ration and having 200 VMs is not a problem.

Initial thoughts are B & E

Could also be B&E

A. The manufacturer and model of the CPUs in the hosts D. The cost to procure and maintain the hardware I considered B to replace D