I have an existing VMware cluster which also hosts my VCSA. I'm trying to add a third host to the cluster, but the VCSA is asking that I put all hosts in the cluster into maintenance mode. The problem is that it involves shutting down all VMs, including my VCSA - ergo, a chicken-egg situation.
Before going too far into this, I thought I would first check in with the community since adding/removing hosts from a cluster feels like an action that shouldn't be so disruptive to operations.
Is it truly the case that adding another ESXi server to the cluster requires that all of the hosts within the same cluster need to be put into maintenance mode, which requires shutting down all VMs? Or am I just choosing the wrong options when adding the new host, and there are options I can select that isn't so disruptive?
If this is the only option, how have you handled my situation where the VCSA is hosted within the same cluster?
Hey @DroboDeveloper ,
Which version on vSphere do you run?
You do NOT need to put ESXi in MM. Simply right-click on the cluster and Add Host...
Next, provide the hostname, root and password, and the rest of the wizard.
If you have vDS, you may want to add a network also:
Right-click on the switch and pick Add and Manage Hosts...
Follow the wizard. Job done.
My blog: AngryAdmin Twitter: @AngrySysOps Youtube: @AngryAdmin
Add the host on the Datacenter level, and move it to the Cluster once added. This way the new host doesn't need to be in Maintenance mode.
I assume that the new host meet the cluster's EVC mode, if enabled.
André.
You don't need to add a host on Datacenter level, as I showed in first reply, you can do it from the cluster level also:
Host doesn't need to be in MM mode
My blog: AngryAdmin Twitter: @AngrySysOps Youtube: @AngryAdmin
There is no need to put a host in a MM. Just follow what @ptarnawski wrote you.
The only obstacle that can avoid adding a host to a cluster are the cluster settings itself, especially EVC mode or HA settings.
One has to check if the new host is sufficient to these settings, otherwise the addition will fail.
I think this is where my problem is...
Out of frustration, I put all of the hosts (except for the one running the VCSA appliance) into maintenance mode this evening. But when I ran the Quickstart wizard (Configure > Configuration > Quickstart) I got a new error message saying:
The operation is not supported on the object. Host XXX.XXX.XXX.XXX running workloads other than VCSA alone or with PSC while undergoing HCI configuration is not supported.
How do I get around this? I just want to be able to do live migration between the hosts and eventually be able to configure vSAN.
1. With latest VCSA we also seen the behavior of a popup msg about MM when adding a Host (with running VMs!) to a existing Cluster. It doesnt matter if that Cluster have EVC enabled or not(all Hosts was same CPU and there is a exact EVC mode available). Anyways.. .our reading was the it referrers to the Hosts we would like to add and not the target cluster.
Our solution was to add the Host on DC Level and than move it into the cluster.
2. @DroboDeveloper now you mentioned "vSAN Quickstart". The solution it not to use the Quickstart function. Just create the HA cluster as normal with up and runing Hosts/VMs (including your VCSA) and than simple "enable vSAN" on that cluster and configure it. So just omit the use of that Quickstart wizzard.
Regards,
Joerg
Disabled EVC and add the host
Hi @RajeevVCP4. If I disable EVC on the cluster, are there any risks of an interruption to the running VMs within the cluster? I want to make sure I do not bring any VMs offline.
@DroboDeveloper No, all VMs will be good. No interruption there.
@IRIX201110141 I'm now wondering whether I'm overlooking something important. Let me further explain...
I have 3 VMware servers. Two hosts are running Intel CPUs; the newest one runs an AMD EPYC.
Historically I was able to vMotion VMs between the Intel hosts without an issue. I'm now noticing EVC is disabled. And so, I assume that it was disabled automatically when I added the AMD host into the cluster (likely a warning message I overlooked). The net impact is that I can no longer do live Migrations within the existing cluster. I'm now wondering if the issue is the mixture of CPU vendors in a single cluster.
I assume this means that I need to create a new "cluster" for the AMD host and accept that live migrations are only available within the Intel cluster? If I want to move VMs between the Intel hosts and the AMD host, will I need to shut down the VM, migrate and bring it back up again?
Sorry if that was a stupid oversight.
You cant mix amd/Intel within a vmotion Cluster.