This post is not as a end-to-end upgrade guide but a methodology guide. Everything is more or less straight forward if one uses the correct methodology.
As this is a home lab I have chosen not to add complexity by adding additional nodes to the various components for High Availability but depend on vSphere HA. This eases and shortens my upgrade path. In a production environment every application should be evaluated and talked through with the stakeholders whether it can rely on vSphere HA or some form of application HA should be introduced.
The release of vSphere 7 introduced a starting point for an update/upgrade round in the lab. I have not always used the methodology when upgrading components as I should have in the lab. When I used this methodology in the current upgrade round it has come to light that some components were not interoperable with each other. Why is this important? If you have a problem with your environment and you call VMware support, they will go through the logs and verify the environment is on par with the documentation.
So where do I start?
Reading the docs first takes time but it will save a lot of time in the long run. Can’t stress it enough RTFM !!
I always create my own high level upgrade guide. This will include the research that I have done and it also includes the upgrade path to follow.
Phase 1: Information gathering
Determine the components and versions
Be thorough in determining the components and versions (BOM or Bill Of Materials).
VMware vSphere 6.7U3
vRealize Log Insight 4.5
vRealize Operations 7.5
VMware Horizon 7.11
You will see below that I had forgotten to include VMware Horizon in my component set. This could have been catastrophic as some components might stop to work when you don’t follow the correct upgrade procedure.
Gather the documentation
After you determine which components are used in the environment you can go and search for the necessary documentation. Find the release notes, upgrade guides and other relevant documentation
What I learned from these documents was that I was not sure the NVMe drives in my hosts would be compatible. After all they are consumer grade NVMe drives and are not on the VMware HCL. I recently installed two NVMe drives, a CT500P1SSD8 and a CT1000P1SSD8 and at the time of install the Crucial CT500P1SSD8 was not recognized. A quick googling showed me a blog post from William Lam that replaced the vSphere 6.7 U3 nvme driver with one from a vSphere 6.5U2 install. I will discuss how I determined if there would be issues around this in ESXi 7.0 in a later phase.
All components should be checked with the vendor and the VMware HCL. Be aware that the vendor and VMware might not always agree and that the VMware HCL might not always be in sync with the the documentation of the hardware vendor. You should always follow the VMware HCL but be aware of the following KB. If vSAN hardware is involved it is advised to use extreme caution as this has a specific section of the VMware HCL for vSAN Ready Nodes and for Build Your Own
VMware Product Interoperability Matrix overview
I did a research on which versions would be compatible with vSphere 6.7 and vSphere 7.0 as I was not yet sure if would be able to upgrade to vSphere 7.0
One product I forgot about was VMware Horizon. I’m currently on 7.10 and the VMware Product Interoperability Matrix show that at least 7.12 is needed to be supported. As I currently use Horizon with a full clone this should not pose much problems and I am planning to upgrade this as well. If I would be using Linked Clones or Instant Clones this could have been worse.
Update sequence for vSphere 7.0 and its compatible VMware products
Now that we checked all the info on whether everything would be compatible and supported after the upgrade, it is time to check the knowledge base article on the update sequence. This article shows what must be upgraded before another component can be upgraded to keep a supported environment.
Phase 2: The upgrade
vRealize Log Insight upgrade
vRealize Log Insight was running version 4.6. Digging back through the release notes showed me that I had to upgrade from 4.6 > 4.7(.1) > 4.8 > 8.1. The VMware Product Interoperability Matrix showed that vRealize Log Insight 8.1 was compatible with either vSphere 6.7U3 and vSphere 7.0
The upgrade process was painless. It just took a lot of time. The process itself is straight forward. Go to Administration > Management > Cluster, upload the pak file and follow the screens. In my case again and again because I did not upgrade Log Insight a long time.
vRealize Operations Manager upgrade
Upgrading the vRealize Operations Manager node is a breeze too, mainly because it is a simple setup with only one master node. vRealize Operations Manager was running version 7.5. I missed the 8.1 release so I upgraded to 8.0 first.
There are a couple of things that need some attention.
Always run the vRealize Operations Manager Pre-Upgrade Readiness Assessment Tool (APUAT pak)
Make sure to upgrade the OS through the OS pak files first, then the vROps pak file
As I upgraded to 8.0 I had to switch files to execute the 8.1 upgrade
The Pre-Upgrade Readiness Assessment Tool showed me warnings for two items:
Validating product version Make sure to run vRealize Operations Manager – 6.6.1, 6.7, 7.0 and 7.5 Virtual Appliance upgrade, as product version is 7.5.0 Ensure product and upgrade versions meet the requirements.
Checking /dev/sda partition size. The size of the partition is less than 20GB. Increase the size of the partition to be greater or equal to 20 GB (https://kb.vmware.com/s/article/75298).
Both are easily addressed. The first one gives a warning to use the correct pak file when upgrading. The second one refers to a KB article that has only a couple of steps:
take vRealize Operations offline
shut down guest OS
increase hard disk in vCenter
boot Virtual Machine
take vRealize Operations online
After addressing both warnings I was able to upgrade.
Simple and easy when you are already on the VCSA with an embedded Platform Services Controller (PSC).
Run the installer and choose upgrade. Supply the source vCenter information and destination vCenter information and click Next – Next – Finish. Grab a drink and wait. It is a two part process. The first part will deploy the new machine with the chosen information and the second part will migrate the data from the old vCenter to the new vCenter.
Before the actual upgrade could take place I needed to be sure that everything would work after the upgrade. Within the vExpert vCommunity I had seen a nice and easy way to do this. I am sorry that I can’t give credits to the person that I got the idea from.
Create a bootable USB ESXi installer or use the iDRAC or equivalent technology to boot your server from the ESXi installer
Find an empty USB flash drive
Put your server in Maintenance Mode
Shutdown and boot your server from the USB installer or the iDRAC or equivalent technology
Install to the empty USB drive – BE CAREFUL not to install to the wrong location
Upgrade check workaround
ESXi will create a VMFS volume from the remaining local space where ESXi is installed by default. After installing I tried to add the ESXi host to vCenter but failed because it had detected the local VMFS volume from the original install and that was conflicting with the one that was still present in vCenter but disconnected. I rebooted the ESXi host, booted into the original drive, verified nothing was on the local drive in the original install and deleted the datastore. Rebooted again into the USB drive and now could add the USB installed ESXi 7.0 to vCenter. Now I was able to get a glimpse of how everything was seen from an ESXi 7.0 install. The NVMe drives I was worried about were showing all fine.
Again this is a home lab and not all components are on the VMware HCL so this adds some extra steps like checking from an actual install. This would not be necessary in a production environment where everything has green checks on the VMware HCL.
The actual upgrade
Upgrading ESXi hosts is done easiest through VMware vSphere – Lifecycle Manager (VMware Update Manager has been rebranded)
I imported the Dell customized ISO, created a baseline and did a Host Compliance Check. The Host Compliance Check was Incompatible and led me to the following two knowledge base articles:
I had to remove everything based on the qedi and qedf drivers
VMware Tools / VM Hardware Compatibility
Upgrading the VMware Tools and the VM Hardware Compatibility is the last part in the process. Determine the viability of each VM to upgrade the VMware Tools and the VM Hardware Compatibility. For most VMs this won’t pose a problem. Nevertheless there are some vendor appliances that will need to run a specific version.
Although vSAN is not really a separate component to upgrade, you will need to upgrade the on-disk format. This is an online upgrade that will not impact the running VMs.
Whilst upgrading the home lab I also decided to rebuild from scratch. There were some challenges to overcome because I have running VMs I don’t want to shut while migrating.
My current home lab setup and the go to setup is documented here (work in progress). Basically it comes down to:
Original setup: three hosts backed with iSCSI storage for running the VMs
New vCenter with two of the three hosts configured for vSAN with connection to the iSCSI datastores
Old vCenter with one remaining host running all of the VMs
Destination setup: new vCenter with vSAN datastore
To migrate the virtual machines from the old environment (from the last remaining host to the two new hosts) I decided to take a look at the ‘Cross vCenter vMotion Utility‘. There is not a lot of documentation available at first sight but it is straightforward to set up and configure. Although I did find some things that are worth noting.
Step 1 : Running the jar
To start the Cross vCenter vMotion Utility one must run a jar file: ‘java -jar xvm-2.6.jar’.
I am running linux (Pop!_OS 18.04) as my OS. I have java version 8 and 11 installed with version 11 as default. Version 11 is not listed on the fling site as supported (Java Runtime Environment 1.8-10: See requirements). Running with version 11 (sudo java -jar xvm-2.6.jar) starts the local website on port 8080 (http://localhost:8080) but does not report back on the CLI.
Under the assumption that the java application started and failed right away, I decided to run it on my windows box which has Java Runtime environment 8 installed. The last line of feedback ‘Initialized controller with empty state’ was the same as on my linux machine. Navigating to localhost:8080 showed the Cross vCenter vMotion Utility web interface. I could now configure the application and run migrations.
It is only later when I closed the running instance on my linux box and restarting it that it showed me output on the CLI that the application started successfully.
Output after restart:
Step 2 : Configuration
Step 3 : Migration
Source Site: source vCenter
Target Site: destination vCenter
Virtual Machine(s): Select one or more virtual machines
Placement Target: Cluster or Host
Network Mapping(s): the utility will detect the source networks for all selected virtual machines and display a selection field for the target network
Storage vMotion does not seem to be supported. I tried to svMotion my machines from their iSCSI based datastores to the newly created vSAN datastore but it failed.
Target Datastore: Shared datastore (same as source)
Choosing ‘Shared datastore (same as source)’ as Target Datastore fails and throws the following error:
I added the destination host and tried again but it also failed with several issues:
destination networks were not listed, only a subset were – although all were added to the distributed vSwitch
matching datastore was not found on the destination host
I could migrate to the new environment but had to select a destination datastore. This posed not much of a problem in my environment because the end goal was to get the virtual machine on the vSAN datastore.
After migrating most of the virtual machines, only two types of virtual machines were left, it felt like I could take a step back if needed. The following types were left to migrate, the vCenter VMs and the firewall VMs. The old vCenter is not needed anymore, the new vCenter and the firewall VMs are and once those are migrated I can go break down the last part of the old setup. The last host will be reset to default settings via the DCUI after which it can be added to the vSAN cluster and I can make the vSAN cluster setup complete. A tmp_vSAN_policy with no redundancy is not the way you (or me) want to run your environment, even if it is a lab environment.
I could not migrate from the old environment to the new environment while also doing a Storage vMotion, I needed to go in steps.
Nevertheless I’m happy to have used the Cross vCenter vMotion Utility. It did save me a lot of work, required little setup and configuration. I didn’t need to change anything to the setup of my old nor my new environment.
VMware I/O Analyzer is a tool to launch orchestrated tests against a storage solution available from the VMware flings website. It can be used as a single appliance where the worker process and the analytics is done within. Additional appliances can be deployed to act as Worker VMs. The Analyzer VM launches IOmeter tests (on the Worker VMs) and after test completion it collects the data. All configuration is done from a web interface on the Analyzer VM.
This post is describing how I deployed VMware I/O Analyzer and how I got to a test with maximized IOs. The first tests were conducted launching a IOmeter from within a virtual machine on the vSAN datastore and showed more or less 300 IOs being generated. In the end 18 Worker VMs with 8 disks each on a 6 host vSAN cluster were used generating 340K+ IOPS. The purpose was to create a baseline for a VSAN datastore maximum IOPs.
6 hosts 1 disk group 1 800GB SSD drive5 1,2 TB 10K SAS vSphere 5.5 U3
The VM OS disks should not be put on the vSAN datastore you want to test, if not the generated IOPs will be part of your report. To keep the Analyser VM IOPS out of the performance graphs, put it on a different datastore.
Deploy one Analyser VM. Deploy a Worker VM per ESXi host. You should end up with as much Worker VMs as you have hosts in your cluster.
I changed the IP of all VMs to static as there was no DHCP server available in the subnet. This means that no DNS entries were required.
Preferably you will want to change the Analyser VM to a static IP as you will manage the solution from a web browser. The Worker VMs you can leave as is if there is DHCP server available. You will need dns entries and change the configuration used here.
To work easily set the Worker VMs on static IPs or create dns aliases as you will be doing a lot of work on the Worker VMs. I prefer static IPs because they add no complexity due to name resolving, etc…
vi/etc/sysconfig/network/routes(The filewill be created ifit doesn’texist)
Add / Change the following line:
(Defaultspace GW space hyphen space hyphen)
Save and close the file (:wq)
Restart the network service:
service network restart
Check if the VM is reachable.
Now shutdown the VM.
Deploying the Worker VM:
Clone the Analyser VM.
Add a Hard Disk of 1GB.
Choose advanced and put the 1GB disk on the VSAN datastore.
I needed to configure static IPs on the Worker VMs, so I had to start each VM and change the IP address. After changing the network settings, shut down the VM and create a new clone. Not changing the IPs will give duplicate IPs.
Ease of access configuration
Two ease of access configurations were applied. The first is configured for easy copying from the Analyzer VM to the Worker VMs. The second because all appliances need to be logged onto for the VMware IO Analyzer solution to work. All commands are executed on the Analyzer VM and then copied to the Worker VMs.
Setup ssh keyless authentication
Generate a key pair
ssh-keygen(with an empty passphrase)
ssh-copy-id will copy your public key to the target machine
The root account password of the destination will need to be supplied for each of the above lines.
BE AWARE: This has the following security downside. If the root account is compromised on the Analyzer vm all worker vms should be considered compromised too.
Change autologon=”” to autologon=”root” in the displaymanager (/etc/sysconfig/displaymanager) file with the following command:
TIP: Create affinity rules in vCenter to keep the Worker VMs on dedicated hosts, otherwise the configuration on the VMware I/O Analyzer dashboard will be outdated soon. The consequence is that certain Worker VMs will not be launching their IOmeter profiles and therefor the reports will not be correct.
Enable the SSH service on the ESXi hosts via the vSphere (Web) Client or through Powershell.
The powershell way: (be aware to filter your hosts if needed). There is a dedicated post about starting and stopping ESXi services through powershell here.
I found that looking at the console of the Worker VMs is interesting for troubleshooting. You can see the IOmeter tests being launched. This was very usefull in the process of creating the IOmeter profile. You don’t need to wait untill the test is finished to see it has failed. Stopping IOmeter tests from the console gives the opportunity to look at, edit and save the launched profile.