Vester – open source, native PowerCLI Configuration Management for vSphere

Recently, I was able to play around a bit with Vester, an open source native PowerCLI configuration management tool for vSphere environments.

The basic idea: 1) Build a configuration file from existing objects (VM, Host, Cluster, etc.), 2) tweak if needed, 3) audit all the things for compliance, and 4) remediate (optional).

At various customer sites, I find myself providing a particular type of code over and over – hey, there’s a hot KB setting tweak we need to apply, so let’s audit current state, test the change, update all objects in your environment, and then leave the customer a script to continue this audit/remediate cycle.  I’d like a better way, following the principle of DRY IT.

While other configuration management tools exist (such as ansible’s VMware module), I find that many VMware admins have a reasonable comfort level with PowerCLI but less so with other CM tools.  Additionally, in my role as a consultant, I may not have scope to engage in a full-blown deployment of CM tools.

However, most of my customers already have a place to run PowerCLI scripts.  This makes Vester an easy, fast solution.

Vester would allow me to hand my customer a PowerCLI module and config script, with simple instructions to run “Invoke-Vester” and “Invoke-Vester -Remediate”. This seems an elegant way to address the above use case. Pretty neat!

Still, in my assessment, this is definitely not a replacement for enterprise “infrastructure as code” and configuration management tools, which provide the feature rich framework required for managing IT at scale.  Instead, it is an easy to understand tool that is accessible to the majority of vSphere admins.  As such, I think it fills a much-needed niche between manual administration and full-blown infrastructure automation.

P.S. I have updated my Resources page to include the beginners guide to “contributing to a github project” I found useful here.  Although I’ve worked with github and some git basics, I found the linked guide helpful to see how the pieces fit together.

vSphere 6.0 PSC Replication Ring Topology

In vSphere 6.0, the PSC introduces multi-master replication.  It is important to understand that the default replication topology is one-to-one, which may not be immediately intuitive.

Design Examples

For example, a site with three sites, each with a single PSC, connected via Enhanced Linked Mode, may follow this replication topology:

In this scenario, Site 1 and Site 3 do not replicate directly, only via proxy with site 2.

Here’s another topology that might “happen”, if you aren’t taking too much care about how you are deploying this:

In this scenario, Site 2 and Site 3 both replicate directly to Site 1, but not directly between Site 2 and Site 3.

In both of these scenarios, all three sites will converge on the same replication set.  However, a single site failure will prevent replication between two other sites.  This design concern becomes more critical in larger environments with more sites.

Recommendation – Use a Ring

For our larger customers we have used a ring topology to improve the design.  Using a ring topology is recommended in the VMware SDDC validated design.  Following our three site example, a ring topology would look like this:

So, how is the replication topology determined?  When you deploy each subsequent PSC after the first, during the install wizard or in the install script you must specify an existing PSC to connect for enhanced linked mode.  A replication relationship will be formed between the target PSC and the PSC you are currently deploying.

How To Deploy a Ring Topology

Ring is simple – it is just a straight line, with the last deployed PSC having an additional replication relationship to the first node in the ring.  Carefully planning the order of PSC deployments will allow you to build the “straight line” replication topology (see the first image).  Once that’s done, log on to the last PSC via SSH and add a new 2-way replication relationship to the first PSC:

For more details, see this incredibly helpful article:

Sysprep Location in Windows vCenter 6.0

Howdy! It’s 2017, and so naturally a good time to load up sysprep files on vCenter 6.0 to support all those Windows Server 2003 deployments!




One of my customers experienced this issue, impacting their ability to use Guest Customization on W2K3 VMs.

    • Default vCenter install on Windows Server 2012 R2
    • C:\ProgramData\VMware\VMware VirtualCenter doesn’t exist, so… after poking around…
    • Placed the W2K3 sysprep files in

Could not customize a Windows 2003 Server.

Note: This was a pre-existing folder.  Also, this path seems akin to the /etc/vmware-vpx/sysprep folder on VCSA, referenced in documentation.  Thus, IMHO, not a crazy decision on the customer’s part to try dumping the files in that folder.


We manually created this path:

After copying the relevant sysprep files into this new directory, as per usual, the vSphere client allowed selecting Guest Customization options while cloning a Windows 2003 VM.