vRealize Automation (vRA) is part of VMware’s Cloud Management Platform and provides comprehensive automation & orchestration capabilities. vRA is available as on-premises installable and available as a cloud service: vRealize Automation Cloud.
The latest version of the on-premises installation of vRA, version 8.1, requires 12 vCPUs and 40 GB of memory. My personal homelab is equiped with a Xeon D-1518 quad core processor, providing a maximum of 8 threads (and thus room for a maximum of 8 vCPUs). The only way to deploy vRealize Automation (vRA) is through the easy installer that leverages the vRealize Suite Life Cycle Manager (vRLCM). The vRLCM will always deploy according to the system requirements (12 vCPUs), and will fail if you don’t have 12 cores/threads available.
Fortunately it’s possible to deploy vRA with less than 12 CPUs, by changing the vCPU configuration after the initial deployment.
This configuration change is only intended for lab/test deployments and will not be supported by VMware.
First, start the easy installer as you would normally do and provide all the required configuration parameters for the vRealize Suite Lifecycle Manager, Identity Manager (vIDM) and vRealize Automation.
I would advice to leave the “Thin Provisioning” option enabled, and also don’t forget the user you configure as the “Default Configuration Admin” in the vIDM configuration step. This user will be granted administrator permissions in vRealize Automation after deployment.
After you’ve configured all the required configuration parameters, the installation commences. First the vRLCM is deployed, vRLCM will deploy vIDM and vRA. You can follow the deployment process on the machine where you started the easy installer, or just check the vCenter Server to see what’s happening.
Installation will fail
After a while you will get an error that the vRealize Automation deployment has failed:
The vSphere WebClient provides more background on the reason behind this error message:
Because 12 vCPUs cannot be scheduled, the VM cannot be powered on. Now change the configuration of your vRA appliance to a configuration that is supported by your (lab) server, in my case 4 vCPUs, and power on the VM. vRA will start, but for the additional configuration steps you will need vRLCM again. So log on to vRLCM and choose “Lifecycle Operations”. Now choose the request in the list that has a red exclamation mark to it:
Choose retry, leave the “deleteVM” option on false and choose Submit.
vRLCM will continue the deployment and run through the vRA post deployment configuration steps. The step “VraVaInitialize” might take some extra time. After a while the post configuration completes:
Now it’s time to logon to vRA 8.1 and continue the setup your Cloud Accounts, Cloud Zones and Blueprints.
I did some first tests and results look promising, vRA is performing quite well and I was able to deploy a first blueprint. The question of course of vRA will do its thing when using some of the advanced features, but for now the platform feels pretty good. For now, happy automating!
2 Comments
corneliunitoi
Hello,
I don’t really understand how this worked for you.
I’ve tried it 3 times already in different ways an I keep getting the same error:
Error Code: LCMVRAVACONFIG590004
First boot check failed on vRealize Automation. Check if all the network details provided are valid.
vRA VA deployment failed at first boot check on host name kovra1.testlab.local
I also installed vRI and vLCM without vRA and then manually deployed vRA with exactly the same result. I’ve also checked DNS and other things people mentioned online and didn’t din issues.
Gabi
viktorious
Hi, on communities.vmware.com I see some comments around your DNS setup. Are you A and PRT records configured correctly?