DefinIT

#vROps 6.5 – Two new enhancements/features I really liked

So vROps 6.5 has been out for around a month now and I finally have a little while to write a post on what I personally really like.

There is a lot of great new enhancements and improvements added in 6.5 but for the sake of brevity I shall highlight two that I really liked.

Flexibility to increase RAM in between node sizes

So if you are familiar with vROps you will know that when you deploy it you have “T-Shirt” sizes to choose from, Small, Medium, Large and now Extra Large (6.5). However an additional extra in 6.5 is support to increase the memory allocation of your vROps nodes, you will not require additional CPU for the node until you hit the next node size bracket. If in doubt check the sizing guide spreadsheet VMware has provided so you can be sure on what you can and cannot do.

This flexibility plus the introduction of a new Extra Large Node size will give you more choice and flexibility with your existing or future vROps deployments.

Direct integration with Log Insight

There was already some integration with Log Insight via the launch in context menu, but now you can access Log Insight from the home page and also when you are troubleshoot individual objects. From a UX perspective this is fantastic and just saves time and further cements the relationship the two products have. You will still need to configure the Log Insight Management Pack and point it to the vLI instance you wish to integrate your vROps cluster with.

 

VM Power On fails with error: cpuid.coresPerSocket must be a number between 1 and 8

| 01/02/2013 | Tags: , , , , , ,

vmware logoHad a strange one after deploying an XP VM from a template today – the VM would not power on and threw the following error:

An error was received from the ESX host while powering on VM [VM name].
cpuid.coresPerSocket must be a number between 1 and 8

cpuid.coresPerSocket Error

Digging around on google the error seemed to be related to over-allocating vCPUs (e.g. assigning 8 vCPUs on a VM with 4 physical CPU cores). This was a single vCPU machine on a 12 processor host, so not likely to be that! It did give me the idea that maybe the VMX had an error, so I edited the VM hardware and added an extra CPU and saved the config. I then edited it back to a single CPU and powered on the machine – it worked!

Examining the vmx showed that the coresPerSocket was set to zero which is incorrect:

cpuid.coresPerSocket = "0"

And after the change, the numvcpus setting was added and coresPerSocket updated:

cpuid.coresPerSocket = "1"
numvcpus = "1"

Fortunately, it’s a simple fix and once I’d updated the template, not something that will bother me again!