V Center HA Feature Fail over scenarios

[zeebahi]

Hi every one.

I am learning new VCA HA  feature , have some questions:

Let say we have  single VCA say VCA1, we have implemented VCA HA for our VCA1.

VCA1 is currently active, while VCA1( cloneVM) is passive.

VCA1 goes down, as a result VCA1 ( clone VM) now becomes active.  After a few minutes,  VCA1 comes back up, what will happen next?

1)Will VCA1 regain its role as ACTIVE or VCA1 ( clone VM) continue to be ACTIVE while VCA1 assumes PaSSIVE role?

2)Will VCA1 assume PASSIVE role because VCA1 ( clone VM) is ACTIVE?

Thanks and have a good weekend!!

[tayfundeger · Accepted solution]

When you activate vCenter HA, a clone is retrieved from the existing vCenter Server Appliance and held passively. If the ESXi Server is down, the witness server activates the passive vCenter Server. The previously active vCenter Server Appliance server now waits passively. If you wish, you can manually activate the vCenter Server Appliance, which is passive.

Note that a minimum of 3 ESXi hosts must be available for vCenter HA.

[zeebahi]

Thanks for helping me understand VM technology, I appreciated it.

If WITNESS  server activates Passive Clone when Active VCA is no longer reachable, then how do we handle a situation where WITNESS node actually goes down while ACTIVE VCA and Passive VCA Clone are still on line.

[tayfundeger]

I don't quite understand your question. Could you give me some details?

[zeebahi]

Sorry about that.

You mentioned:

When you activate vCenter HA, a clone is retrieved from the existing vCenter Server Appliance and held passively. If the ESXi Server is down, the witness server activates the passive vCenter Server

Based on the above, It is witness server which controls when  Cloned VM will become active if existing vCenter goes down.

Let say we have a following situation;

Existing Vcenter is still up and running,  Cloned VM is passive but for some reasons, witness server is down.  How  does Vcenter HA behave if witness server itself crashes?

Hope it clears up.

Thanks.

[tayfundeger]

If the Witness node fails, the Active node continues to function and replication between Active and Passive node continues, but no failover can occur. If the Witness node is repaired or comes online, it automatically rejoins the cluster and the cluster state is healthy.