Upgrade NSX-T from 3.2.1 to 4.0.0.1

In this blog I am doing a walkthrough of upgrading nsx-t from 3.2.1 to 4.0.0.1.

Before you start make sure DRS is enabled and it is full automated & that you have shared storage for your virtual machines in your cluster. As the upgrade needs to put hosts into maintenance mode.

Download the upgrade bundle NSX 4.0.0.1 Upgrade mub file from the vmware site.

https://customerconnect.vmware.com/downloads/details?downloadGroup=NSX-4001&productId=1324&rPId=92330

Do a backup of your current nsx-t configuration

click system > click backup and restore > click start backup

Clear any nsx-t manager alerts

Click home > click alarms

I had password expiration alerts, so I used the following commands on my edges and managers to update them.

ssh to your managers and edges

set user admin password

set user root password

set user audit password

if you want to remove the password expiration from the nsx-t accounts use the following command

clear user admin password-expiration

clear user root password-expiration

clear user audit password-expiration

In a vcf environment use “set user admin password-expiration 9999” as no expiration can lead to problems in a sddc manager environment. I would not setup expiration 9999 or remove expiration in production, but a lab should be fine

Click system > click upgrade > click upgrade

Select upload MUB file > click browse

Select the .mub upgrade file > click upload

Click prepare for upgrade

Accept agreement > click agree and continue

Run pre-checks

Click the drop down on run pre checks > click all pre-checks

Click RUN PRE CHECKS

Fix any issues that the precheck shows
Click next

Select Serial > Select after each group completes > select edge group > click start

If edge upgrade fails reboot the affected edge and take it out of maintenance mode

After the reboot, the version has changed

Retry the update > click continue

Once completed check your edge nodes in nsx-t manager. system > fabric > nodes > edge transport nodes
One of my edge nodes has a mismatch > Click mismatch

Click vSphere/Edge appliance > Click resolve

Go back to the upgrade and click next

On the host upgrade page > click serial > click when an upgrade unit fails to upgrade > click after each group completes > select the cluster/s you want to upgrade > click start

You will notice in vCenter it will start putting esxi hosts in maintenance mode.

As the upgrades continues the esxi version will update in nsx-t manager

Click system > click fabric > click nodes > click host transport nodes

Check the nsx version column

One of the hosts failed the update and needed to be rebooted as resolve did not work

Once rebooted resolve worked and it updated the version

Click nsx install failed > click resolve

Select the error tick box > click resolve

Click resolve

Update completed

Go back to nsx-t manager upgrade page > click on hosts tab

All successful now

Click next

On NSX Manager tab click ”2 issues needs acknowledgement”

Click acknowledge if your happy with the warnings > click ok

Click start

Click confirm to start the upgrade now > click start (this will do the upgrade of the nsx manager appliances one at a time & but you might lose access to the nsx-manager)

nsx-t manager failed to be put into maintenance mode

Use the get cluster status command to get the id should be in the error message
reboot the failed nsx-t manager
Retry update
Click GOT IT (message is saying nsx-t manager won’t be available during upgrade and reboot)

You can click on the NSX Manager appliance page to check if the update has completed

Click system > click Appliances

Both nsx-t managers have been updated now in the appliances page

If you get a message during the upgrade of nsx-t manager saying it cannot enter or exit from maintenance mode reboot the nsx-t manager in question and re-run the upgrade.