What we check
We analyse the components of your infrastructure and assess them against three yardsticks: the vendors’ best practice, our own Good & Common Practices, and experience from other estates.
On top of that come the questions that arise in day-to-day operation. Where can the work be done more efficiently, where should security be tightened, what is more awkward to manage today than it needs to be.
With what
For VMware by Broadcom we use three tools.
VCF Checker is an HTML health report for VMware Cloud Foundation 9 and vSphere 9. It is based on vCheck-vSphere by Alan Renouf, stays under his MIT licence, and soulTec funds its continued development. Around 110 plugins check vCenter events and appliance health, cluster HA, DRS and compliance, host hardware and configuration, datastores, network security, snapshots, VMware Tools and hardening. The VCF Operations integration is optional. The report contains only the sections where there is something to do. It needs PowerShell 7.6 and VCF.PowerCLI 9.1, so it also runs on macOS and Linux. Source: github.com/virtualFrog/vChecker.
Invar reads the inventory out of vCenter and exports it in RVTools’ format, also across platforms. Source: github.com/virtualFrog/Invar.
We are allowed to use the VMware Health Analyzer as a partner. It reads the data out of vSphere and Horizon and compares it against the vendor’s best-practice catalogue.
For HPE we use InfoSight. It evaluates usage data automatically and checks it against known problems and security gaps, on the servers through the Active Health System and iLO, on the storage through the arrays’ telemetry.
For particular areas we have written our own configuration checkers. Those are not published.
Versions and compatibility
For VMware by Broadcom we check against the HCL and the Broadcom Compatibility Guide to see whether the components in use are supported. We check against the interoperability matrix to see whether the combination of versions in use is supported together.
The firmware and driver levels sit there too. A host can consist entirely of supported parts and still fall outside the matrix in combination.
What we cover for it
Six categories: storage systems, SAN, backup, servers, server virtualization and client virtualization.
| Area | Systems |
|---|---|
| Storage | HPE Alletra, Primera, Nimble, 3PAR, MSA · IBM StorWize |
| SAN | Brocade SAN switches |
| Backup | Veeam Backup & Replication · IBM Spectrum Protect |
| Servers | HPE ProLiant, Apollo, Synergy, C7000 · Lenovo ThinkSystem |
| Virtualization | VMware vSphere, meaning ESXi and vCenter |
| Client virtualization | Omnissa Horizon, App Volumes, DEM and UAG |
Older systems are listed on purpose. We also check estates that are not on current releases.
False positives
Some of the findings do not apply to your estate. Most of those come out of the Health Analyzer.
Reports about PCI placement inside a host are usually beside the point.
When a tool suggests spreading the virtual machines differently across the cluster, it often does not know the architecture underneath. On a stretched cluster the distribution it objects to is the intended one.
RemoteDisplay.maxConnections limits how many console sessions to a VM may be open at once.
Up to vSphere 7 the default was 0, from vSphere 8.0 it is -1. The report flagged -1 as a
deviation from the standard even though the value switches the limit off.
There is no general list of these cases. They depend on what the tool does not know about the architecture or the version. So an engineer who knows the estate goes through the findings.
What you get
The report, usually within a week. We present it, as a summary for a leadership audience or in detail with the technical team. You get it as a PDF with a task list and the responsibilities.
You can do the work yourself or together with us. That is a separate engagement.
