Which action can resolve the issue?

Refer to the Exhibit.

Users are having difficulty accessing a Web server since a new Web application was configured in a virtual machine running on sc-titanium03.vmeduc.com. The web server has been assigned 1 vCPU. The vSphere Client reports the error shown in the exhibit. DRS is set to fully automated mode, but the problem has not resolved.

Which action can resolve the issue?

Set a CPU reservation for the web server VM.

Manually move the Virtual Machine using high priority vMotion.

Configure the DRS Migration Threshold to Aggressive.

Set CPU affinity for the virtual machine.

7 Comments on “Which action can resolve the issue?

  1. Tony says:

    Why “A”? We are not sure that it will solve the problem. What if there is not much CPU left to be reserved on the host due to other reservations? The host CPU is busy, so we need to offload the VM to another host.

    My guess is that the VM is configured for host affinity (in this case to load balance web servers), so DRS is not moving the VM. Thus, v-motioning the VM to the other host could solve the issue until the host CPU issue is resolved.

    Unfortunately, this is another poor question from VMWare.

  2. Gt says:

    Other vms are taking too much cpu. By setting reservation to this web server, it will solve the issue. So A is correct

  3. Pablo says:

    A) Set a CPU reservation for the web server VM: there is not enough resources for everyone. By reserving CPU for the web server I assure it access to the physical CPU that is missing and I solve the problem. Yes, the other VM’s will have now to share the rest and I may have problems with them but I have resolved the problems the question asked.

    As for the others, they are clearly False:

    B) Manually move the Virtual Machine using high priority vMotion: makes nothing. If DRS didn’t migrate the web server (since there is no mention to it, you assume DRS works fine and is well configured) means that the cluster is in contention, and there is not enough resources for the VM in any of the two host that makes the cluster.

    C) Configure the DRS Migration Threshold to Aggressive. that just means that any small improve DRS could find migrating a VM, will be done. But with an alert like the one shown, less than agressive threshold will normally migrate the VM

    D) Set CPU affinity for the virtual machine: the stupid answer easy to rule out 🙂

  4. Rich says:

    This question seems to require one of two assumptions:

    1. That the names of the VMs indicate that they are load balanced VMs and have affinity rules set that prevent both servers of the same type to run on the same host. In that case neither B or C would work. D is nonsense, so that leaves A as the only possible choice. Considering the alarm, A is a poor choice, but if you make this assumption then it’s the only possible choice.

    2. If you do not make the assumption above, the other assumption would be that the DRS migration threshold is set to the most conservative setting which only automatically migrates when the host put in maintenance mode. Because answer C says to set it to ‘Aggressive’ rather than ‘to a more aggressive setting’, I’d rule this out. Setting the migration threshold to fully aggressive will cause the VMs to bounce too much.

    So I’d go with answer A.

    In reality the question should ask ‘What settings should you investigate to find the reason why DRS is not migrating the VM.” The answer would be ‘The DRS Migration Threshold setting’ and ‘The DRS affinity rules’.

    Would anybody really just set the CPU reservation based just on what’s in the exhibit, without investigating the situation further? I doubt it.

Leave a Reply

Your email address will not be published. Required fields are marked *