Every vSphere admin meets the Managed Object Browser eventually. Usually it’s 11pm, a support article says “open the MOB”, and you discover a corner of vCenter that looks like it was designed in 2004. (It roughly was.)
It’s worth knowing properly, though. The MOB is one of the most direct ways into your VMware estate, which is exactly why it’s useful, and exactly why hardening guides want it switched off.
What is the VMware Managed Object Browser?
The Managed Object Browser, or MOB, is a web interface built into vCenter Server and ESXi. It shows the vSphere API object model as plain HTML pages: every managed object (hosts, virtual machines, datastores, clusters, networks), its properties, and the methods you can call on it.
You reach it at:
https://<vcenter-or-esxi-host>/mob
You log in with vSphere credentials, land on the ServiceInstance, and click your way down through content, rootFolder, datacenters, and so on. Nothing is hidden behind a friendly UI. What you see is what the API sees.
What is it actually used for?
Mostly three things:
- Finding managed object reference IDs. Every object has a MoRef, like
vm-1234orhost-56. Scripts, support cases and third-party tools often ask for one, and the MOB is the quickest place to look it up. - Inspecting properties the vSphere Client doesn’t show. Runtime state, extension registrations, and configuration buried a few levels deep are all there in raw form.
- Invoking methods. This is the one to be careful with. From the MOB you can call API methods directly, for example to unregister a stale plug-in extension that’s cluttering the vSphere Client. VMware support articles sometimes walk you through exactly that.
That third point is why the MOB isn’t a harmless viewer. Anything the logged-in account is allowed to do through the API, it can do through the MOB, with no confirmation dialogue and no change ticket.
How do I enable or disable the MOB?
On ESXi, the MOB has been disabled by default since ESXi 6.0. It’s controlled by a host advanced setting:
Config.HostAgent.plugins.solo.enableMob
You can change it in the vSphere Client under the host’s Configure → System → Advanced System Settings, or with PowerCLI:
# Check the current value
Get-VMHost "esxi01.example.local" |
Get-AdvancedSetting -Name Config.HostAgent.plugins.solo.enableMob
# Turn it off again when you're done
Get-VMHost "esxi01.example.local" |
Get-AdvancedSetting -Name Config.HostAgent.plugins.solo.enableMob |
Set-AdvancedSetting -Value $false -Confirm:$false
On vCenter Server, the MOB is available by default. It can be turned off with the enableDebugBrowse option in the vCenter service configuration. Check the current guidance for your vCenter version before you change it, as the method varies between releases.
Should the MOB be switched off?
On ESXi hosts, yes. VMware’s security configuration guidance and the DISA STIGs for ESXi both require the MOB to be disabled, because it exposes the full API surface and can be used to change configuration or gather information about the host.
The sensible pattern is simple:
- Leave it off by default.
- Turn it on for a specific task, on the specific host that needs it.
- Turn it off again as soon as the task is done, and check that it actually is.
The last step is the one that slips. Someone enables the MOB for a support case on a Friday, the case closes, and the setting stays on for the next two years. Your compliance scan finds it at the next audit, if you’re lucky.
What should I use instead?
For anything you’ll do more than once, use a tool that talks to the same API but leaves a better trail:
- PowerCLI for Windows-centric teams and most day-to-day automation.
- govc for quick command-line inspection.
govc object.collectgives you much of what the MOB shows, in a form you can script and diff. - pyVmomi when you’re building tooling of your own.
These are repeatable, reviewable and easy to log. Clicking through the MOB is none of those.
Keeping VMware estates in line
The MOB is a small example of a bigger pattern. VMware estates are full of settings that are fine to change briefly and risky to leave changed: SSH enabled for a fix, a lockdown mode exception, a debug option nobody reverted. Each one is easy to correct. Finding them all, continuously, across every host, is the hard part.
That’s the kind of work autonomous cloud operations is built for: detecting drift from an approved baseline and putting it back, with a full audit record, rather than waiting for the next quarterly review to notice.