HA Insufficient resources

[sc_21111]

Hello all,

I got a strange issue with a new 2-node cluster that has only two virtual machines, one per physical node, that are stored on local storage, so they cannot migrate.

The nodes have 384GB Ram each.

The "local" VMs that cannot migrate has 72GB RAM all reserved.

So even considering both the Vms the total requested RAM would be 150GB 

There's the Vcenter VM which has 16GB Ram.

So the great total would be 176GB which is less then 384GB that each node has.

I don't understand why the Admission Control is warning about insufficient resources

[Duncan Epping]

If they are stored on local storage why would you even have HA enabled? HA requires shared storage to begin with?

[sc_21111]

Hello, obviously this is not the final situation where we're going to add vm on shared storage.

the question is why the HA is warning for insufficient resources when is seems not to be the case.

the vcenter is already on the shared storage provided by the two HPE appliances that virtualize the storage.

[sc_21111]

Hello Duncan,

 obviously this is not the final situation where we're going to add vm on shared storage.

the question is why the HA is warning for insufficient resources when is seems not to be the case.

the vcenter is already on the shared storage provided by the two HPE appliances that virtualize the storage.

[a_p_]

With only 3 VMs (vCSA and two VMs) and sufficient resources on a two node cluster, I'd consider disabling Admission Contol.

Anyway, with what you wrote, the warning could be caused by the Adminssion Control configuration, and the reservations that you set for the VMs. Please explain exactly how Admission Control is currently configured, how many VMs you run in the cluster, and what the highest reservation (RAM, CPU) is for the VMs.

André

[sc_21111]

Hello

the admission control is set as "standard" with cluster resource and host fail to tolerate 1

at the moment there's only the vcenter, which can migrate, and the two SDS appliances which cannot migrate as they're local to each server.

each of those VMs has 72GB of reserved memory.

[a_p_]

From the documentation:

"The basis for vSphere HA admission control is how many host failures your cluster is allowed to tolerate and still guarantee failover."

If in a two node cluster one of the hosts fail, there will be no other host available to which the VMs can failover anymore. Thus the "insufficient resources" message.

André

[sc_21111]

Hello André

I found this note that explains it 

Note: The cluster resources percentage option for admission control also checks that there are at least two vSphere HA-enabled hosts in the cluster (excluding hosts that are entering maintenance mode). If there is only one vSphere HA-enabled host, an operation is not allowed, even if there is a sufficient percentage of resources available. The reason for this extra check is that vSphere HA cannot perform failover if there is only a single host in the cluster.

thanksfor you input

[Casey]

Hi sc_21111, 

With the VM's on local storage, you can not utilize the HA portion to restart on another host without some crazy configurations. However, you can enable HA to restart the VM's in the situation of a power failure. 

The issue you're encountering with the Admission Control warning about insufficient resources can be attributed to how vSphere calculates available resources and admission control policies. Here are a few points to consider:

1. Admission Control Policy

vSphere's Admission Control is designed to ensure that there are sufficient resources for failover in the event of a host failure. By default, it reserves enough resources to accommodate the failure of one or more hosts, depending on the policy you have set.

2. Host Resource Reservations

Even though each node has 384GB of RAM, the reserved resources for the VMs and the overhead required by the ESXi hosts might cause the Admission Control to trigger a warning.

3. HA Slot Calculation

vSphere HA calculates the number of slots available on each host based on the highest CPU and memory reservations of any VM in the cluster. A slot is a logical representation of the resources required by a VM.

4. Resource Overhead

Each VM and the vCenter VM have some overhead associated with their memory usage. This overhead is part of the resources that need to be considered.

Steps to Resolve the Issue:

1. Check HA Settings:

   • Go to Cluster Settings -> vSphere HA -> Admission Control and review the policies. Consider setting a custom failover capacity or adjusting the percentage of cluster resources reserved as failover spare capacity.

2. Review VM Reservations:

   • Ensure that the memory reservations for the VMs are not excessively high. The reservations can be found under VM Settings -> Resources -> Memory.

3. Review Host Resource Usage:

   • Check the resource usage on the hosts to ensure they have adequate free resources. This includes the overhead for the VMs and the ESXi host.

4. Adjust Slot Size:

   • If you are using the "Host Failures Cluster Tolerates" policy, the slot size can be a limiting factor. You might need to adjust the slot size or switch to the percentage-based policy.

5. Cluster Configuration:

   • Verify the cluster's total resource pool and compare it to the required reservations and overhead. This might help in understanding if there are any misconfigurations.

Example:

Here's a step-by-step guide to check and adjust HA settings:

1. Access Cluster Settings:

   • Open the vSphere Client and navigate to the cluster.
   • Click on Configure -> vSphere Availability -> Edit.

2. Adjust Admission Control:

   • Under Admission Control, choose Percentage of Cluster Resources Reserved.
   • Set a percentage that reflects the actual requirement, considering you have sufficient resources to failover VMs in case of a host failure.

3. Monitor Resource Usage:

   • Go to Hosts and Clusters, select each host, and review the Resource Allocation tab to check the memory and CPU usage.

4. Review VM Configurations:

   • For each VM, right-click and select Edit Settings.
   • Go to the Resources tab and check the Memory and CPU reservations.

By carefully reviewing and adjusting these settings, you should be able to resolve the Admission Control warnings and ensure your cluster operates without issues.

------------------------------
Please mark as Recommended if this solved your problem. This will help the community find solutions to their problems as well.

MORE LOGS, MORE SCREENSHOTS, MORE DETAILS = Easier To Help You :D

I have super ODC, so excuse my excessive formatting.
------------------------------------------------------------
VCP-DCV 8.x | VCP-DCV 6.x | CompTIA A+, Net+, Sec+, Proj+, Server+ | MCSE | MCSA | MCITP
------------------------------

[Jimmy15]

"The "local" VMs that cannot migrate has 72GB RAM all reserved."
Remove reservation, It will reduce the SLOT size.