Upgrading to VxRail 7.0.0

As I mentioned in this post, we launched VxRail 7.0.0 today (April 28).  In order to get the upgrade package, you will need to visit SolVe Online to get the link for the upgrade package.  This upgrade package will handle not only upgrading your cluster (check the release notes for upgrade paths if you don’t want to see them in SolVe) to vSphere 7.0.0 as well as handle embedding the PSC for customers that have deployed the VxRail provided vCenter.

Before you start, I can’t stress enough to reference SolVe for your cluster (and for each cluster) to make sure your cluster can do the upgrade.  Also, this is a rather large upgrade and therefore will take some time (for those with the VxRail provided vCenter, we have to deploy the new VCSA and then embed the PSC and migrate the database info from both the legacy VCSA and PSC).

This upgrade itself is pretty much on par with the previous upgrades.  Once you get the link to the upgrade package, you need to upload the package:

After the package has been uploaded, you’ll see the summary screen with the components that will be updated and their versions:

As you can see in the screenshot above, there’s quite a few hardware updates in addition to the software updates.  If you have any GPUs in the nodes, please make sure you have the appropriate driver/firmware package for those nodes.

After clicking on “Continue Update”, you will be prompted by the system with the traditional information requests for usernames and passwords.  The biggest difference here will be for our customers with the VxRail provided vCenter, the system will ask for a temporary IP address.  This IP will be used to deploy the new 7.0 VCSA with an embedded PSC.

After you provide the credentials (and temp IP, if required), the system will then start the upgrade on the cluster.  Remember, this is a completely non-disruptive upgrade.

One thing some customer may have noticed was that in 4.7.300, we were able to bring in some optimization to the upgrade process.  In the past, we would wait for a node to completely finish its upgrade before going onto the next node.  We’re now able to run some of the operations in parallel on 2 hosts at a time to improve the upgrade timing.  Please note, that even with this change, only 1 node will ever be in maintenance mode or rebooting at any given time, which means we will keep the VMs and data available during that time.

If you have the VxRail provided vCenter running on your cluster, during the upgrade, you’ll see the 7.0 VCSA get deployed using the temporary IP you provided.

The system will continue with upgrading each node until it’s fully completed.  After that, you should have an completely upgraded environment to VxRail 7.0.0.  For those customers with the VxRail provided vCenter, you’ll also see that we rename your 6.7 VCSA and PSCs.  When you’re ready, you could remove that.

That’s pretty much it!  There’s obviously a little more to it, which is why  you really need to make sure you go to SolVe Online to get the rest of the upgrade details.

One last thing I did want to highlight.  With this release, we’ve also put up guardrails to make sure our customers are able to stay within a supported config.  As I mentioned in this post, we’ve been able to prevent customers from using vLCM to upgrade their VxRail cluster.  This will help make sure customers stay on our fully validated, end to end stack for their cluster.  If you try to use vLCM, you will get an error similar to this:

This is mainly to make sure that our customers stay in a fully supported configuration and don’t run into any issues applying any updates that may not have been validated with VxRail.