During that discussion, I do talk about the fact that you can scale out or scale up the vROps cluster depending upon the requirement. While Scale Out is simple and you can do it by adding new nodes to the cluster, scale up can be trivial in some cases.
While working with a customer, where we decided to scale up the vROps nodes from Medium to Large, I found out that the small and medium vROps appliances when deployed for the first time are deployed with hardware version 7. This means that the appliances cannot go beyond 8 vCPUs as hardware version 7 has a configuration maximum of 8 vCPU for a virtual machine. This means that when you try to upgrade from a Small or Medium node to a Large vROps node, you will hit a wall unless you upgrade the hardware version of the appliance.
In this case, while you can update the hardware version, please ensure that you do not touch the VMware tools version. In other words you can update the hardware version to move from 8 vCPUs to 16 vCPUs without any issues. Needless to say that the hardware version upgrade would include a downtime. Just to be clear on the sequence of steps:
- 1- Make sure you have a Full Backup of your appliances ( a general recommendation before hardware version upgrade).
- Bring the vROps Cluster offline from the Admin UI.
- Shut down the appliance(s) gracefully using the vSphere Client.
- Right click on the appliance(s) and chose the option of upgrading the hardware version. I would encourage you do this using vSphere Web Client to get to the latest supported version.
- Make the nodes of large configuration i.e. from 8 to 16 vCPUs and 32 GB to 48 GB.
- Power on the appliance(s).
- Bring the cluster back online.
This process will automatically increment all memory configurations for Java Processes associated with Cassandra, Collector and other services within the application configuration files.
Please ensure that you to stick with the VMware recommended (tested) sizes of Small / Medium / Large as per the following KB article
Share & Spread the knowledge!!