Hi!
In the vSphere client, I choose to view the cluster summary page.
What does the following mean:
current memory failover capacity 75%
current cpu failover capacity 87%
configured failover capacity 25%
We are using HA
Regards
H
VMware HA is configured with a min % of resources free guarantee (in your case 25%).
Actually you are above this limit both for CPU and Mem.
For for info see: http://www.vmware.com/pdf/vsphere4/r41/vsp_41_availability.pdf
How do you have your HA set up ? I can be set up to tolerate a sertain ammount of node failures , or a % of resource reserved
by the looks of it you have 25% of your resources configured for HA ( ie 1 node out of a 4 node cluster ? )
VMware HA is configured with a min % of resources free guarantee (in your case 25%).
Actually you are above this limit both for CPU and Mem.
For for info see: http://www.vmware.com/pdf/vsphere4/r41/vsp_41_availability.pdf
Hi again,
Yes, the cluster is set up with 25% failover capacity.
But what does memory failover capacity 75% really mean? Are we currently using 75% of our total memory in the cluster or what?
Thanks for helping out!
Regards
H
It's a failover capacity.
It means that you have 75% of free resources
-----
I write wrong. To see how the failover capacity is calculated and related with resourse usage see:
http://www.yellow-bricks.com/vmware-high-availability-deepdiv/
There is a section specific for the "Percentage of cluster resources reserved" case
Sounds strange, because all the host is using more then 50% of there memory. So how can 75% of the memory be free? I know that VMware can share memory between machines (we have a lot of Windows machines), maybe thats why?
The 75% value does not mean you have it free on your cluster. Means that 75% of your hosts can fail (in the RAM perspective) and the vCenter still will be able to perform the failover. This obviously makes sense because of the overcommitment technologies on the VMware products (ballooning, memory compression, etc).
In fact, you need to take the smaller % number to know how many ESX can fail at the same time on your cluster.
Also, you configured it to 25% failover capacity as minimum. This means that if any of the % goes to this point, no other VMs will be powered on at your cluster.
"In fact, you need to take the smaller % number to know how many ESX can fail at the same time on your cluster."
What do you mean by that? What smaller number?
In your example:
current memory failover capacity 75%
current cpu failover capacity 87%
The smaller % is 75. So you can have 75% of your hosts failing at the same time in this scenario.