IIn this lab, we will do a number of tasks.
- Create a cluster Enabled for vSphere HA
- Add an ESXi host to a cluster
- Test vSphere HA Functionality
- View the vSphere HA Cluster Resource Usage
- Manage vSphere HA Slot Size
- Configure a vSphere HA Cluster with Strict Admission Control
Task 1: Create a Cluster-Enabled for vSphere HA
First, we need to right-click on the Training datacenter and select ‘New Cluster’ and then it needs to be configured to the lab. Now it did tell me to select a policy but there was no option for it, as the book is using an older version of vSphere.
Task 2: Add your ESXi Host to a Cluster.
For this, I just need to drag my ESXi Host into the cluster.
I also need to add my other host as well.
Now it does get confusing because we should be doing both students tasks in the lab so I actually need two VMs under ESXi 2 as well so I need to register them as well, but I can’t because they are in the process of going into high availability. So I had to put the ESXi 2 into maintenance mode to move it, now I can add some VMs under it and then move it back.
That’s the one thing I don’t like about this book is because it gets confusing to remember to double every task we are doing.
Anyway here is my Lab Cluster now.
And this is the summary of the vSphere HA, it says that I should be able to see my slave hostname, but I can’t and I assume that’s a version thing.
Question: Does the number of protected VMs match the number of VMs in the cluster? Yes it does.
Question: How many datastores are used for heart beating? Just the VMFS storage.
Now looking at the configuration issues we can see it’s complaining about not having network redundancy which we need to address.
To fix this we need to go into the hosts vMotion Kernal Adapter and edit the settings.
But first it turns out I need a separate VMFS datastore for my ESXi 2 so I need to set that up.
I’m going to have to do a lot of juggling for space
This is actually a process…. I’m not getting the heartbeat datastore issue going away.
I might have to re-add the host for it to rescan it???? I’ll try it. I could just delete all the VMS and restore my storage and add it half and half for each host, that’s kind of last resort though.
Okay, I fixed the issue, I had to move all my hosts back to the folder, delete the Cluster and then remake it and put the hosts back.
Now I can go back to my vMotion VMKernal adapter and edit the settings, making sure to do this on both hosts.
now I need to reconfigure the Hosts for vSphere HA.
Question: Did the management configuration error disappear? Yes, it did the only message I have now is about the gateway not being seen which is understandable because I don’t have one working.
Task 3: Test vSphere HA Functionality
Now I need to go to the hosts VMs and make note of its current VMs
Now I need to simulate a host failure by rebooting my master ESXi.
And then go into the Lab Cluster monitor tab and then look at the events.
So I got a massive amount of errors says the host does not have enough CPU and memory to handle the swap over, so I’m going to have to adjust that.
I’m having to shut down all the VMs on host 2 and then shut the host down to perform the upgrade.
Actually what I might do is just shut down all but one machine, that will still prove that failover works.
So now I will restart my master host and see if the VM changes over.
So I am having troubles with vMotion HA working an I found that my times on my ESXi hosts were way off so I had to set the NTP server again and I had it set to the wrong server, so i set it to my DC and bam it’s fixed.
Now let’s try this again…
It finally worked.
All the Master host VM’s came over, the one that was powered up took a lot longer than the one that was shutdown but that makes sense.
Question: Has the master host changed? No, it has not.
Task 4: View the vSphere HA Cluster Resource Usage.
To monitor the resource usage we can go into Lab Cluster and resource reservation.
And making sure the VMs have no reservations.
And making sure no reservations are set on the VMs
Task 5: Manage vSphere HA Slot Size
For this, I need to go into vSphere HA in Lab Cluster and edit it.
Now I face another version issue with this book… I am suppose to edit the Admission Control and select Define host failover by a static number of hosts but all I see is the following.
I’m going to select ‘Dedicated failover hosts’ and hope it does not bite me in the ass because let’s think logically here. A static number of hosts would be the same thing as Dedicated failover hosts? I just have to select my hosts I want to use now.
Now I have added them I’ll click Ok and progress on.
Again it’s saying, in summary, I should be able to see slot cluster information and this is all I see…
It seems to be because I am using the wrong Admission Policy.
Ok, I found out what the menu is called, it’s called slot policy now, now I can see my runtime info.
Now that has been sorted, I can go over to VM’s and Templates and set a reservation for CPU on nw##-3
And I can’t forget about the other VM.
Now I have to see if my Advanced Runtime Info changed, and yes it has.
Now we can use HA slot policy to enforce a slot size.
Going into edit vSphere HA and Admission Control I need to edit the CPU slot size to 300MHz
Now I can click calculate which allows me to view the link.
Now I need to record my required slots for my nw##-3 VMs which is 2.
Now I need to view the slot information for this cluster, here w can see the slot size has once again changed.
Now we have to remove the vSphere HA fixed slot size setting, which is done by changing the slot size policy back to Cover all Powered-on VMs
Now we can remove the CPU reservation from our VMs
Let’s not forget about the other one.
Task 6: Configure a vSphere HA Cluster with Strict Adminssion Control
Before I can do anything I need to shut down all my VMs
After waiting for the memory to be released I can look at the Lab Cluster summary annd record its information.
Question time! Why is the free memory value less than the total memory capacity value?
This is because there is VM’s still dependent on the host’s memory, and because there are still functions such as vMotion in the background that memory is reserved for.
Now I can assign memory to a VM through Resource Reservation.
Now I need to record the Advanced Runtime info for Total slots in the cluster.
Question Time! Why does the vSphere Web Client report that value?
This is because none of the VMs are currently active.
Now I need to power on a VM and refresh the report, and answer some questions.
Question: How many slots in the cluster, used slots, available slots, and failover slots do you see?
Question: Why is the value for the failovers lots half the number of slots?
This is so the host can tolerate the failover.
Slot size values.
Question: How is the memory slot size calculated?
The memory slot is calculated by the highest memory reservation.
Now I Need to power on the other VM on the same host and refresh the Advanced Runtime Info.
Question: How many slots are available and for what reason?
There are now only 3 slots as I have just powered on another VM.
Question Can you power on another VM? Yes I can because I still have free slots available
Question: If a cluster has N total slots, can you still power on N VMs? No, because some a left for failover, also this question does not make much sense.
Task 7: Prepare for Upcoming Labs
Now I need to remove my memory allocation on my VMs.
Now I need to remove my Lab Folder from the inventory
And then I need to allow VMs to exceed failover capacity.
This lab was HUGE!!!
It taught me a lot about failover and how to manage it, but my biggest issue was my damn size allocation for my VMFS that set me back so much, after that the lab was okay besides the documentation being outdated.
Note to self, don’t over do your storage allocation!!