VMware Cloud Community
GeneNZ
Enthusiast
Enthusiast

Increasing available slots causes "insufficient resources" warning in HA?

Hi there,

I'm wondering if anybody could help me understand my issue with slot sizes and HA. I've read through multiple blogs in an effort to understand exactly how slot sizes work and how they apply to an HA cluster, but from what I'm reading, doesn't seem to match up with our virtual environment. A most helpful blog site by the way was: .

Firstly I'll start with my problem. I received an "Insufficient resources to satisfy HA Failover in cluster XXXX in XXXX Datacenter" warning after having moved two VM's into the a cluster, bringing the VM count from 16 to the present 18. Now my understanding from reading the blogs, is that I would receive this warning when the resources required by the VM's outnumber the available slot sizes, if one host were to go down. The Current Failover Capacity should also drop to 0 hosts. Problem is that I get the above warning and my Current Failover Capacity remains at 1.

The configuration of our environment is as follows:

  • 2 x vSphere4 Servers identically configured with 32GB Ram and 25Ghz CPU for a total cluster size of 64GB Ram and 50Ghz CPU. CPU utilization is around 10% and Memory utilization is around 20-30%.

  • vCenter 4

  • 18 VM's with no individual CPU or Memory reservations (i.e. their all set to 0 for their reservations). The maximum memory overhead is 175mb and we have 2 machines with 2 vCPU's and the rest have 1 vCPU.

  • Admission Control is enabled, and the Current Failover Capacity is 1 host.

With the default 256Mhz/256MB for das.vmCpuMinMhz and das.vmMemoryMinMB variables I receive the following from the Advanced Runtime Info (this configuration gives the error):

Slot Size: 517Mhz, 2 vCPU's, 431mb

Total slots in cluster: 94

Used slots: 18

Available slots: 29

Total powered on vms in cluster: 18

Total hosts in cluster: 2

Total good hosts in cluster: 2

Now what confuses me from the above stats is that there are more than enough available slots for the virtual machines.

I decided to play around, and adjusted the das.vmCpuMinMhz and das.vmMemoryMinMB advanced HA variables. I noticed that if I adjusted the values to 256Mhz/512mb respectively, the warning message disappears with the following Advanced Runtime Info:

Slot Size: 517Mhz, 2 vCPUs, 689MB

Total slots in cluster: 84

Used slots: 18

Avaiable slots: 24

Total powered on vms in cluster: 18

Total hosts in cluster: 2

Total good hosts in cluster: 2

What confuses me here is increasing the minimum memory decreases the slots, and therefore it should make things worse, not better?

Can anyone with a better understanding, help me clear up my confusion.

Thanks in advance.

Gene

Reply
0 Kudos
1 Reply
GeneNZ
Enthusiast
Enthusiast

Hi,

I've worked out that changing the resource pool CPU and Memory requirements (by lowering them), and restoring the minimum memory and cpu values to 256mhz/256mb removes the 'insufficient resources' warning. The next question is how do resource pools affect the HA slot calculation in vSphere4?. Note I am using "Host failures cluster tolerates" and not "Percentage of cluster resources reserved as failover spare capacity".

Also apologies for the subject title that doesn't make sense. I double checked the contents but not the title.

Gene

Reply
0 Kudos