An administrator observes the following symptoms for a virtual machine:
CPU usage is consistently above 90%
CPU ready value is consistently above 20%.
Application performance is impacted.
Which two actions should the administrator take to improve the performance of this virtual machine? (Choose two.)
A.
Increase the number of vCPUs assigned to this virtual machine.
B.
Decrease the number of vCPUs assigned to this virtual machine.
C.
Verify that VMware Tools is installed on every virtual machine on the host.
D.
Increase the CPU shares assigned to the virtual machine.
Explanation:
Explanation/Reference:
CPU usage is consistently above 90%
CPU ready value is consistently above 20%.
Those are typical symptoms of low CPU shares. D is correct.
You need VMware Tools to collect information from guest operating systems to ESXi hosts. C is correct.
http://pubs.vmware.com/vsphere-60/index.jsp#com.vmware.vsphere.html.hostclient.doc/GUID-28C39A00-743B-4222-B697-6632E94A8E72.html
possibly, B,C and D are right.
http://pubs.vmware.com/vsphere-60/index.jsp#com.vmware.vsphere.monitoring.doc/GUID-C5984D55-7E84-4F2F-855A-AF637D340DEC.html
Verify that VMware Tools is installed on every virtual machine on the host.
…
Increase the CPU shares to give the virtual machine more opportunities to run.
…
Reduce the number of virtual CPUs on a virtual machine to only the number required to execute the workload.
How to choose two of three???
Alex, I also think B, C, & D are correct. This question was pulled directly from the link below and it indicates all three of them are correct. The second link is a very similar type of KB and it suggests B & D. D is obviously correct so it comes down to either B or C. If this were a real world problem and D didn’t solve the problem then the next thing I’d do is B. C doesn’t seem to have a lot of benefits except that it is a VMware best practice. So I guess I’d say that B & D are the best answers.
http://pubs.vmware.com/vsphere-60/index.jsp#com.vmware.vsphere.monitoring.doc/GUID-5F8147A1-6416-4D29-BA3D-E4CED3966016.html
https://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=2001003
As soon as I posted this I realized why C makes sense. If this is actually a memory issue then C would allow ballooning and other memory management to occur which could resolve the issue.
Whatever. This could be B, C, or D IMO. Going back to the link that this question is drawn from, C is the top item listed under the solutions followed by D. I don’t know that B can be ruled out in any case.
It could be B, If the host is really loaded and shares are out of line removing vCPU’s from the VM is going to have little effect. Choose the 2 options that will deliver the best results first. C, D. Fix best practice first, re-configure you CPU shares then lower the vCPU’s.
A and D are correct, since the CPU contention happened at guest OS level, not host.
%RDY is the esxtop CPU counter and shows how much time a VM was waiting to be scheduled by the hypervisor for execution. So it’s NOT about a cpu contention in the guest.
I am still researching this one. If this was an MS cert question, the answers would be what would be administratively easier. I think that rules out C due to the potential work involved in checking each VM and the possible subsequent reboot (ie outage) you would need to coordinate.
Based on the link from above, http://pubs.vmware.com/vsphere-60/index.jsp#com.vmware.vsphere.monitoring.doc/GUID-5F8147A1-6416-4D29-BA3D-E4CED3966016.html
C is still valid though it’s not a simple fix.
Chris, it is actually very easy to check if all VMs have Tools installed, just select the host and select the VM view and check Tools are installed/version column, select to view if not visible.
Unfortunately, it’s not clear where – on ESX or within the VM – 90% CPU utilization is observed…
Not ‘A’ – because we already have very high CPU Ready % – the sign of contention on ESX host. Adding more vCPU would only aggravate the issue and increase this number.
Not ‘B’ – purely from semantics of the question, this may already be a VM with 1 vCPU only, thus no further to decrease 😉 Besides, assuming 90% relates to VM’s vCPU utilization, making it less SMP for sure would not help.
‘C’ – makes perfect sense if 90% CPU utilization is to be attributed to ESX host (which can be due to heavy memory swapping) because it would allow for taking advantage of various memory saving techniques like ballooning and page sharing.
‘D’ – obvious choice under choice under circumstances.
A and D as its within the VM symptoms not ESXi and I deal with this daily on my 100K VMs daily, globally.
CD
http://pubs.vmware.com/vsphere-60/index.jsp#com.vmware.vsphere.monitoring.doc/GUID-5F8147A1-6416-4D29-BA3D-E4CED3966016.html
So while in practice I would agree with Dude they’ve taken this question straight from the help pages.
How I hate this VCP exam. If you ask a guy with photographic memory who’s never heard of VMware to do this exam he would pass it with ease..
VMWare is disqualifying this exam all by itself, you can find 80-90% of the answers on the documentation pages. If you are an admin that is able to find the documentation pages..you Sir, are a VCP!
working link: https://docs.vmware.com/en/VMware-vSphere/6.0/com.vmware.vsphere.monitoring.doc/GUID-5F8147A1-6416-4D29-BA3D-E4CED3966016.html
This is a poor question and I kind of agree. The information is freely out there and there’s not much real-world stuff being asked. But, it is what it is.
There are three correct answers, but based on the above link, you want to FOLLOW THE ORDER OF THE PROCEDURE. First available answer on this question is ensuring VMTools are installed, second available answer on this question is increasing the shares, then decreasing the vCPUs.