My vSphere 7.0 Upgrade story - step by step!

I started working with vSphere 7 when it was RC and it was pretty cool.  So I was excited to get it into my lab for real use.  Even though I do not yet have a vCenter license from vExpert yet.  Here is my upgrade story - it is not the best one I have shared, but it ends well.
I really like the new layout of the VM summary screen. I am curious about the single image management too.
4/11/20 - I see they released the upgrade best practices now - https://kb.vmware.com/s/article/78205. Good info to check out before you upgrade.

Dependencies

  • Horizon View - Check VMware interoperability for View, and I see that 7.12 is supported and I have done that upgrade very recently so no problem here.
  • Veeam Backup & Replication - I checked Veeam’s web site to confirm if it supports vSphere 7 or not. Looks like not.  But I confirmed it will work with 7. I confirmed with Veeam R&D that it was not an accident that it worked, but rather they did work to make sure when it shipped it would work when vSphere 7 dropped. (BTW here is a patch that makes the vSphere 7 support official).
  • Runecast - is not yet supporting vSphere 7 but very soon.  So I will unregister it’s plug-in so that while I upgrade it doesn’t get hurt.
  • If you have any other apps, like the NetApp HCI VMware Plug-in, even when it is supported by vSphere 7 it is a good idea to unregister the plug-in before the upgrade and register it again afterwards. My guide on this is updates are fine to not unregister but upgrades is really good to unregister.  One of the pre-upgrade checks even looks for extensions and says they will not survive.  I saw that again when I forgot to unregister the NetApp HCI VMware plug-in.

So now we have checked our dependencies we can proceed.  View and Veeam will work after I have upgraded and it won’t be a long wait for Runecast. Mind you, I really hope not since I have only 10 days left on my license!

Upgrade vCenter

Note1: While I normally use a Mac to run the upgrade process it was not so good this time. A lot of security prompts where some I could get by, and one (ovftool) I could not get by.  So did this in Windows. I am reusing some of the screenshots I got on the Mac to save time.

Note2: Make sure that SSH is enabled on your vCSA as it will fail a pre-upgrade check if it isn’t and your upgrade will not work.

Note3: I like to make sure there is no outstanding updates before I start the upgrade.
Note4: Please do an image backup of your vCSA before you start.  Just in case.  Be aware you cannot do image backups of a vSphere 7 vCSA.
This will take a bit. It is like an install but we use the upgrade button.  You will need to know a few things: root and administrator account to vC, as well as a host root password to hold the new vC. This is a migration type upgrade after all. There is an outage in vC for a bit but it will not impact applications running in VMs.
  • You mount the ISO for the vCSA.
  • Now look for the vcsa-ui-installer folder

  • Now choose the OS you are working with.
  • In this article I started with a Mac and continued with Windows as I mentioned above.

  • Now look for the installer file.

  • The process starts.

  • We are doing an upgrade so we use the Upgrade button!

  • We have an info screen that you can read before hitting the Next button.
  • Next is an EULA screen so just check the box and hit Next.
  • We now need to connect to our source vCSA.  The one we are upgrading (migrating).

  • After we fill in the FQDN of the vCSA we use the Connect button.

  • We need to fill in a few passwords for the source appliance, and I always fill in the info for the ESXi host who is holding the vCSA and I think that is the best way to do this rather than the existing vCenter.

  • In the next screen, seen above, we fill in the ESXi host info for where we will put the new vCSA.  I always choose the host that is NOT holding the current vCSA.

  • Now we select deployment size.  The default of Tiny is good for me lab but choose what makes sense to your environment. Getting this right now is MUCH easier than trying to fix later.
  • On the next screen select your datastore.
  • On the next screen we need to do the network settings correctly for the new vCenter appliance.

  • You need to use a temporary IP for the appliance, and it will be changed to the proper IP once the original vCSA is powered off.
  • On the next screen you will see we have successfully deployed the vCSA and that we need to continue with Stage 2. There may be a slight wait for this screen.

  • We see we have done Stage 1, and need to do Stage 2. That is the upgrade part, and the import part.

  • Now we see pre-upgrade checks happening.

  • I get an error as SSH is not enabled on my vCSA which is why up above it says to do that!

  • I get some warnings from the pre-upgrade check.

  • There is no surprises in the two you see above.  There is a third I scrolled down for which was a warning about my extensions not working after the upgrade.
  • When you Close, you get to make a decision.

  • I always choose the last option, as it gets more info to the new appliance.  Which may be good for me, or maybe good for support.
  • The next screen is about being in the CEIP which I strongly recommend you do as it will help you out in the long run! None of your secrets are sent to the cloud.
  • Next we see a summary screen.

  • You have to confirm you did a backup of the vCSA before you started and hit Finish.
  • I will show a few screenshots below as things progress.

  • Once the process is done you may seem some messages displayed - I did.

Now we do a little test.

So that is good.  Can I log in now with my AD credentials?

So the upgrade (migration) when well.  Now to deal with my hosts!

In a few days, when there has been no issues with the new vCSA, you can delete the old vCSA.  Make sure it is not turned on.

Upgrade ESXi Hosts

We are upgrading our hosts from 6.7.0 Build 15160138 to 7.0.0 Build 15843807.

Note1: I like to make sure there is no outstanding updates for the hosts before I start the upgrade.
Note2: Based on previous upgrades, I will put each host, one at a time, into Maintenance Mode and restart it.  This will make things go smoother.
Note3: Based on previous upgrades, I will make sure vMotion works and that hosts can get into and out of maintenance mode.
Lets get started.
  • Change to the LifeCycle Manager area.

  • Add the ESXi 7 ISO in the Imported ISO’s area.

  • Use the  button to get our vSphere 7 ISO loaded.

  • Now we need to create a baseline for this new OS.
  • Change to the Baselines tab.

  • We will use the New button to create a new baseline for vSphere 7.

  • On the next screen select the vSphere 7 ISO.  There should only be that one choice in fact.
  • On the next screen is a summary. Hit Finish.
  • Now change to the Hosts & Clusters view.
  • Select your cluster.
  • Then change to the Updates tab.
  • You will see mention of single image to manage things.  That is new and I will not use it at this time. I will use the sort of traditional method to upgrade, and will look at single image for adding my other hosts.  Look for an article on that.

  • Use the Attach button we see above to add our vSphere 7 baseline.
  • Use the Check Compliance button we see above to see what is reported.
  • We see what we expect to see.

  • Everything looks good except the vSphere 7 baseline is not compliant.  Lets fix that!
  • Select the baseline and you will be able to select Remediate.
  • Approve the EULA and you will see the screen below.

  • You can expand the options at the bottom of the screen for more options. I do not see the do multiple hosts at the same time option. I really like the option for new clusters to move things along faster.
  • When ready hit the Remediate button.

  • I should mention on the first cluster I upgrade (this is the second cluster) it failed right away with an error of cannot get exclusive access.  So I restarted the hosts and did the upgrade of them one host at a time.  That worked good.
  • I don’t see anything happening, so I click on the Monitor tab of the cluster.  Then look at the Tasks view.  I select the Remediate entity task and look below to see what it has to say.

  • Not sure what is happening as there is no error messages. Oh wait, now it says hosts cannot enter maintenance mode. Which is odd as I tested that, and vMotion before I started!! So I check the esxupdate.log file in /var/log on the hosts to see if it gives me any info - and it does not. This is frustrating as I checked maintenance mode and vMotion before I started so really confused why the patching process cannot do that.
  • When this sort of thing happens, I move to update one host.  So I am putting one host in MM and will update it and see what happens.

  • So you can see above it looks very similar.  Select the vSphere 7 baseline and Remediate.
  • At the host level, change to the Monitor tab, and select Tasks and watch the Remediate entity one. In my case I can see 1 hosts are being remediated. And it finishes and is updated.
  • So now we do the second host. And it works too.  So both hosts are now at 7.0 and all the VMs are fine so that is good.  But we are not done yet.

VM Updates

Once the hosts are upgraded, you now work with VMs. Important to note that there will be restarts of the VMs and in fact likely multiple so there is going to be outages here.
  • Start at the cluster level, and select Updates, followed by VM Hardware.

  • The first thing you do is use the Check Status option. This is quite helpful as it shows the info on each VM including if there is an upgrade available or not.
  • In my case all the VMs had a status of Upgrade Available so I selected all of them and choose Upgrade to Match Host.
  • W get an info message.

  • It is good to know that there is reboots so there will be a short outage, and only five can be updated per host at a time.
  • I notice it did not select nor will it update, my templates.  That will be up to me to do manual next time I update them.
  • Ok, lets do this. We want to get this whole thing done right?!
  • When it finished I still had a few VMs not upgraded so I selected them and did the upgrade again.  I did not worry that my vCSA was not done. I left it as it was.
After that use the VMware Tools baseline area in updates.  It behaved oddly in one of my upgrades.  Even though the status said finished it took four or five more minutes after that for the Tools install to be finished. So lets be careful this time. You can in this UI also set some VMs to auto update if that works for you.
  • The first thing to do again is Check Status.

  • Now we select the VMs to upgrade.  There are some to not select.  Like our vCSA, or appliances like Log Insight.  But select the rest.  A good guide is any that say Guest Managed do not upgrade.
  • One we have selected the VMs, and used the Upgrade to Match Host option, we see a similar dialog.

  • I like how it gives you a chance to deselect if that is necessary.  Remember this process can take four or five minutes longer than it is suggested in the UI.  You can check on the VM to see if it has the current version of VMware Tools.
  • This baseline is problematic as far as I am concerned.  It does most of the VMs, but not all.  Then I do those.
Once this is all done we test.

Testing

  • Can you do a backup? Might need to do refresh connection before you try due to new cert in the VCSA.
  • Can you do a restore?
  • Can you connect to a desktop?
  • Can you log in to the VCSA using your AD credentials?

Updates

  • 5/29/20 - Veeam has released a patch for vSphere 7 official support and other fixes.
  • 4/11/20 - the vSphere 7.0 Upgrade Best Practices are now out - https://kb.vmware.com/s/article/78205.
  • 4/9/20 - the issue with patching occurred in my work lab today.  So the workaround of patching at the host is what saved the day.
  • 4/5/20 - I have thought about it, and I think it might be best to not use the VMware Tools baseline but rather update VMware Tools the way you do it now.  Will be more consistent I think.

Conclusion

I am glad we made it this far.  I am very sorry I could not upgrade my entire cluster at the same time.  I do not get support, so looking at the log files and trying again is the best I can do. I have two more hosts to bring into this cluster, and I am going to look at the Single Image stuff to do that.  I Will write it up of course.

Questions and comments always welcome.

Michael

=== END ===

 

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.