VMware – Read before upgrade to vSphere 6.5

VMware

Yesterday VMware announced the general availability of the brand new vSphere 6.5. They announced the new version at this years VMworld in Barcelona. But now you can download and install the bits. But there is a catch. Please make sure you read and understand all the important information before upgrading to vSphere 6.5 because there might be some limitation at the moment. Let me bring some light into the darkness.

Compatibility considerations

You should not upgrade to vSphere 6.5 if you are running one (or some / all) of these software components in your environment:

  • VMware NSX
  • VMware Integrated OpenStack
  • vCloud Director for Service Providers
  • vRealize Infrastructure Navigator
  • App Volumes
  • Horizon Air Hybrid-Mode
  • Integrated OpenStack
  • vCloud Networking and Security
  • vRealize Business for Cloud
  • vRealize Configuration Manager
  • vRealize Hyperic
  • vRealize Networking Insight

These components are not yet compatible with vSphere 6.5. But as we know VMware, they are already working for updates. Please check the VMware Product Interoperability Matrix for further information about updates to the products above.

  • If you have to revert a migration, please check VMware KB2146453 for reverting a vCenter Server to Appliance migration.
  • To roll back a vCenter Server Instance on Windows, please check the vSphere Upgrade Guide.

Upgrade Considerations

Before upgrading your environment, review these critical KB articles to make sure the upgrade will be successful.

vCenter Server

vCenter Server to vCenter Server Appliance

PSC High Availability

ESXi

NSX

vRealize Operations

vSphere Web Client

Known Issues

vCenter Server

vRealize Operations Manager

Security Considerations

TLS protocols

Encryption considerations

  • Running a encrypted KMS virtual machine can cause a loss of data in the event of a host failure.

More details in the VMware Knowledgebase (KB2147548):

https://kb.vmware.com/kb/2147548

*** Update ***

Backup Considerations

There is one thing i missed to mention. If you are using Veeam Availability Suite v9.5 then you can’t do backups with vSphere 6.5 at the moment, because Veeam does not support this vSphere version yet. But also the guys at Veeam are working on an update, which will be (historically) release about two months after general availability of the new vSphere version.

So stay tuned!

VMware vSphere – How to script vMotion for your VMs

vMotion Script

VMware vMotion is a pretty good feature regarding the availability and load balancing in your vSphere environment. Today i created a vMotion script to help me create a backup with a backup software.

As so many times my blog posts are the result of a problem i had and for which i needed a solution. It shouldn’t be different today. I worked in my vSphere homelab. I created some virtual machines and installed my backup software of choice. My idea was to have a backup before doing any work with the VMs, just in case i screw it up. So i can easily go back to a known good state of the VM and try again. But this task wasn’t so easy.

As a vExpert, VMCE, MVP, Trainer or many other different tech people you can request a NFR license key for Veeam Availability Suite. So did i. The NFR key was delivered quickly to my mailbox, and was even faster installed in Veeam. But there was a catch. At least my NFR license is limited to two sockets, but with no limits for protected VMs, and it comes with a full 1-year retention period, instead of just 30 days as the regular trial.

So i had to deal with the fact that only one host (i’ve got three hosts in my lab with two sockets each) is protected by Veeam. This limitation woke the hunter in me because i had to find a solution. My goal was to backup all my VMs but with only two licensed sockets. The approach I chose was to set vSphere DRS to manual, then do a vMotion of all VMs to the host which helds the Veeam license, doing a backup and set DRS back to fully automated after backup. If you are working with ressource pools you shouldn’t disable DRS, because that results in removing the ressource pools. But there is a workaround for that too. Instead of creating a new problem i did the easy way and just set DRS to manual.

How to get the vMotion script

If you’re familiar with GitHub you can download my script from there:

https://github.com/driftar/vSphere

For any other user i’ll provide the script directly here:

[code lang=”powershell” gutter=”true”]
# .SYNOPSIS
# This script will start a vMotion of all virtual machines on a specified datastore to a specified ESXi host.
# If you are working with a backup software which is licensed to a specific host, # this will probably help you.
# Only recommended in smaller environments or if you have enough ressources on this host.

# .DESCRIPTION
# The script loads a PSSnapin; it sets some PowerCLI options; it connects to your vCenter Server with the given credentials;
# it gets all your VMs in an array; it starts then a Host vMotion of all the VMs in the array to a specified ESXi host.

# .NOTES
# File Name : pre-backup.ps1
# Version: : 1.0
# Author : Karl Widmer (info@)
# Prerequisite : PowerShell V2 over Vista and upper / VMware PowerCLI 6
# Tested on: : Windows Server 2012 R2
# with PowerCLI : PowerCLI 6.3 Release 1 build 3737840
# with PowerShell: 4.0
# Copyright 2016 – Karl Widmer / driftar’s Blog (www.)

# .LINK
# Script posted over: https://www.driftar.ch

# Load PowerCLI cmdlets
Add-PSSnapin VMware.VimAutomation.Core -ErrorAction "SilentlyContinue"

# Set PowerCLI behaviour regarding invalid certificates and deprecation warnings
Set-PowerCLIConfiguration -InvalidCertificateAction ignore -DisplayDeprecationWarnings:$false -confirm:$false

# Define vCenter User and target Datastore
$vcHost = ‘vcenter.domain.com’
$vcUser = ‘administrator@domain.com’
$vcPass = ‘password’
$datastore = ‘your_datastore’
$cluster = ‘your_cluster’
$targetHost = Get-VMHost -Name yourhost.domain.com

# Connect to vCenter
Connect-VIServer $vcHost -User $vcUser -Password $vcPass

# Get VMs (pass array of VMs to $VMs, for example ‘get-datastore test | get-vm’)
$VMs = Get-Datastore $datastore | get-vm

# Get Cluster information to set DRS to Manual for backup window
Set-Cluster $cluster -DrsAutomationLevel Manual -Confirm:$false

Foreach($vm in $vms) {
Write-Host ("Start Host vMotion for VM ‘" + $VM.Name + "’")

Move-VM -VM (Get-VM -Name $vm) -Destination (Get-Vmhost $targethost) -RunAsync

Write-Host ("Waiting…")

Write-Host ("Host vMotion for VM ‘" + $VM.Name + "’ finished")
}

# This last script step should probably be executed in a post-backup script step.
# It sets the DRS automation level back to fully automated. Your VMs will then probably load-balance on your hosts.

# Set DRS on cluster back to FullyAutomated after backup window
Set-Cluster $cluster -DrsAutomationLevel FullyAutomated -Confirm:$false
[/code]

Update 07.11.2016

After updating my ESXi hosts to 6.0.0 Build 4510822 my script stopped working. So i simplified the script and released version 2.0.

VMware vSphere – How to find floppy images (paravirtual SCSI)

Paravirtual

Today’s blog post is about some tricks and treat. It’s soon Halloween, folks… Today i’m working in my vSphere homelab. I tried to create a virtual machine. Well, that wasn’t that complicated. More complex was to find the floppy images for providing the PVSCSI driver. You’ll need this driver when you’re creating a paravirtual SCSI storage controller within your virtual machine. A quick search provided a good VMware knowledgebase article which helped me. Gathering the floppy images was then easy as pie.

You can’t access the “vmimages” folders through storage browser, it shows to be empty. This is an expected behaviour because the vmimages search through vCenter Server is turned off in ESXi 4.0/4.1. And i just assume that it’s also turned off in newer ESXi versions, like 6.0 which i’m using in my homelab. Because i actually can’t find the “vmimages” folder. For my case i installed all my nested ESXi hosts from scratch, manually, so there shouldn’t be any issue. This should also happen in production environments.

But now lets start to gather this folder. How do you get your hands on it? Let me show you.

Start putty and connect to one of your ESXi hosts (SSH service has to be enabled in the security profile / firewall on your ESXi host).

Run these two commands to create a folder on a datastore where you whish to store the images:

[sourcecode language=”plain” gutter=”true”]mkdir /vmfs/volumes/<datastore_name>/floppies
mkdir /vmfs/volumes/<datastore_name>/vmtools[/sourcecode]

Where <datastore_name> is the name of a shared datastore.

Run these commands to copy the contents of the original source folder to the newly created folder on the shared datastore:

[sourcecode language=”plain” gutter=”true”]cp /productLocker/floppies/*.flp /vmfs/volumes/<datastore_name>/floppies/
cp /productLocker/vmtools/*.iso /vmfs/volumes/<datastore_name>/vmtools/[/sourcecode]

Now you’re able to mount the floppy images in your VM to get access to the paravirtual SCSI driver.

Happy virtual machine creation, and Happy Halloween too 😉

Halloween