HA with insufficent capacity in cluster

[LOPERA]

We had a HA Cluster with two Host 5.1, all VM run in the first host (12 VMs), and vCenter Appliance Linux.

When we disconnected of the energy the first host (shut down simulation), the second host only restart 4 VMs (including the vCenter Virtual Appliance), and the vCenter show the message "Insufficient capacity in cluster HA_CLUSTER1 to satisfy resource configuration in MY_DATACENTER"

We had configured the HA with "Specify failover host" and don't work, we try with "percentage of cluster resources: 50%" but, is the same error.

We has only two host  in the Cluster.

Thanks in advance

[jdptechnc]

When your first host goes unavailable, you would see the "insufficient resources to satifiy HA" message regardless, because, at that point in time, you do not have the resources to satify the failure of a host (you only have one host at that point).

Look at the events tab in vCenter for the VM's that did not start up, are there any clues there?

[LOPERA]

Hi, thanks four you reponse,

In the task events for the VMs thar dis not start, there is nothing, but in the DATACENTER task there are this errors messages:

Do you have any recomendation?

[aravinds3107]

Welcome to the Communities..

"Insufficient capacity in cluster HA_CLUSTER1 to satisfy resource configuration in MY_DATACENTER"

The above error mean there is no enough resource (CPU/Memory) avilable on other host in the cluster to failover the other VM's

Does both your host have same amount of CPU & Memory?

Is there any reservation set for any of the VM's?

Is there any priorities set for the VM failover?

Also check Diagnosing insufficient fail over capacity on a VMware High Availability Cluster

[LOPERA]

Thank you for the welcome!

As Jdptechnc said "When your first host goes unavailable, you would see the "insufficient  resources to satifiy HA" message regardless, because, at that point in  time, you do not have the resources to satify the failure of a host (you  only have one host at that point".

I understand that this is normal in this point, but not all VMs restart in the other host, only 4 VMs, I think that the VMs that did not start are blocked in this moment, I don't now.

I used in the Admision control the 'Specify failover host', this means that the other hosrt is only as failover or backup host. This host is identical to the other, men, cpu, datastores, all (see the pic).

I deactivate the Admision control option and it is the same issue, only 4 VMs restarted! (see the pic)

This the priorities in my last test (see pic), before I had some VMs with High and other with intermediate priorities.

Thanks.

[admin]

Can you confirm that the virtual machines that did not fail over are actually compatible with srvesx2 (the healthy host)? A quick way to check is to manually power one of these vms on the healthy host. If that succeeds, can you check the fdm.log file on the healthy host for failover errors (grep for "FailoverAction")

Elisha

[jdptechnc]

Were you CPU or Memory overcommitted prior to the first host going unavailable? Did you have these VM's already powered on before enabling HA admission control? Maybe you were already at a point where you were beyond your failover capabilities, but since your existing VMs were already powered on, it wasn't evident until you tried to fail over?

When you have both hosts online and all VM's powered on, what does vSphere HA Cluster Status tell you?

Also, did you check to see if the VM's are actually capable of running on the second host?  Are all the VM networks and datastores used available on both hosts?

[aravinds3107]

Are the VM's which failing to failove is showing as Protected in the VM Summay tab?

[LOPERA]

Thanks.

As Eszikind say, I found in the FDM.log this message: /vmfs/volumes/bc12754c-d04

ded91/SRVPROD-DC/SRVPROD-DC.vmx, vim.fault.NoCompatibleHostI found in the Vmware documentation the solution: http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=203699...I have some machines with the ISO connect to a local Host datastore, I diconnected all CD-Roms of the VMs.
Tomorrow I will do the first test, I think that it's the problem.Thanks.

[LOPERA]

Hi, we resolved the problem. All test was ok.

Thanks.

[arunachalam]

Thanks...