Upgrading vRA 7 to vRA 7.2

Upgrading vRA 7 to vRA 7.2

 

 

So I thought I would update my vRA 7.0.1 installation to the newly released 7.2.

I found the VMware upgrade document here and noticed there are some pre-requisites that must be in place before you perform the upgrade:

VMware Requirement

Minimum Required

My Appliance

Minimum RAM requirement

18GB

18GB

Minimum COU requirement

4 vCPUs

4 vCPUs

Minimum Disk 1 size

50GB

18GB

Minimum Disk 3 size

25GB

25GB

Minimum Disk 4 size

50GB

50GB

IaaS Web Server/Model Manager - Java version

Java 8 update 91 or higher

7 Update 67

MS SQL Server - Java version

Java 8 update 91 or higher

8 Update 111

IaaS Web Server/Model Manager - .Net version

4.5.2

4.5.1

MS SQL Server - .Net version

4.5.2

4.6.1

IaaS Web Server/Model Manager – Free disk space

5GB Available

20GB

MS SQL Server – Free disk space

5GB Available

30GB

 

The good thing is if your platform meets the minimum requirements you can go ahead and perform an in place upgrade. If like me you need to perform some remediation work most of the tasks are pretty straight forward.

On my lab installation, I am only using a single appliance and all of the IaaS components are installed on a single server, so it should go without saying that on a distributed environment there will be further steps to take.

Prerequisite Configuration

Java Upgrade 

Make sure when you are upgrading your Java version you are installing the x64 bit version, the version that automatically downloads from the java website appears to be 32bit, so I had to grab the offline x64 version. When I ran through the Java upgrade the installer prompted me to uninstall the older version of Java that was installed which was handy.

Make sure there is an environment variable for JAVA_HOME:

.Net Upgrade

I hadn’t run Windows Updates for a while on the IaaS server, I let that run through and the .Net updates were downloaded and then installed.

Resizing Disk 1 on the vRA appliance

OK so the most involved piece of the remediation work was actually resizing Hard Disk 1. There is a step by step guide on the VMware site here I thought I would run through it and point out any gotcha’s I find.

First off we need a spare Linux virtual machine, I am using a CentOS 7 machine, which also has the desktop GUI installed.

It should go without saying that in a production environment you need to backup all of your servers before you begin this, as I don’t currently have a backup solution in my lab I took a clone of the vRA7 appliance, and I took a snapshot of the IaaS Server and a backup of the SQL Server.

OK, so power off your vRA appliance, and increase the size of Hard Disk 1 to 50GB:

Then you need to add the vRA appliance primary disk to the spare Linux machine:

Log in to the CentOS machine and check the disk utility:

We can see that the vRA disk that has been added to the Linux is device “sdb”, so we need to use the disk partition management for the device:

Now to list the existing partitions type “p”:

You’ll need to take note of the partition table, as we’ll need the information a little later on.

Now we delete the existing partitions note: nothing is committed here until we write the changes
Type “d” and select partition 1 then do the same for partition 2:

Create Primary Partition

We need to create the new partitions now, type “n” and then “p” and choose partition 1. OK so this is where we need to check back with the partition table we printed off before.

Looking at the existing partition table above, we can see that the primary partition starts at block 2048, so select that as the first sector. For the last sector we need to do a quick calculation as you can see the default option would use up all available space, but we need to make sure we leave enough space for the swap partition, so what we do is work out the amount we need to reserve:

Swap Partition

End Block – 37748735
Start Block – 31438848

37748735 – 31438848 = 6309887

Available size – 104857599
Swap size – 6309887

104857599 – 6309887 = 98547712 (I rounded down to 98547711)

So the last sector figure we need is 98547711:

Now we make the partition bootable by typing “a” and selecting partition 1:

Create Swap Partition

Type “n” and “p” and choose partition 2, you can select the default starting and ending sectors here also:

Complete the process

Now we need to assign the partition code ID to each partition, again we get this from the original partition table:

Type “t” and then select the ID relevant for each partition (83 for the primary partition, 82 for the swap partition)

Before we commit the changes we should check the new partition table:

Providing you are happy with the new configuration write the changes and exit the partition tool by type “w”

Let’s format the swap partition by typing “mkswap /dev/sdb2”:

Now clean the primary partition by typing “e2fsck –f /dev/sdb1”: