Home / Spherifyer / ESXi hardening
Check ESXi hardening, host by host, without scripts.
SSH switched on for troubleshooting and forgotten, a host without a time server, a VM with an emulated network adapter: nobody notices such deviations day to day. After every collection Spherifyer checks all hosts and VMs against items from the vSphere Security Configuration Guide and reports what deviates, with a recommendation.
How to check ESXi hardening by hand.
In the vSphere Client per host under Configure: in the services whether SSH and ESXi Shell are running, the lockdown mode in the security profile, the time configuration, and syslog target and shell timeout in the advanced system settings. With PowerCLI it works for all hosts at once, for example:
Get-VMHost | Get-AdvancedSetting -Name Syslog.global.logHost, UserVars.ESXiShellInteractiveTimeOut
Then the VMs: hardware version, network adapters, connected drives, VMware Tools. And the result is only valid for the day of the check.
- Every host
- Services, lockdown mode, time and logging sit in different places and must be checked host by host.
- Every VM
- Hardware version and network adapters are visible per VM, not in one shared list.
- Drift
- SSH is switched on for maintenance and stays on, a new host joins without a syslog target.
- Evidence
- Audits do not ask about today’s state but whether the requirements are met continuously.
Every deviation as a warning with a recommendation.
Spherifyer reads vCenter at the configured interval, every hour by default, and then checks all hosts and VMs. Whatever deviates from the vSphere Security Configuration Guide appears among the warnings in the Compliance category, each with a suggestion of what to do. The categories let you filter the list down to exactly that.
Spherifyer only needs a user with the “Read-only” role and never changes anything itself. You apply the recommendation, and Spherifyer sees on the next run whether the warning is gone.
- ESXi services
- SSH or ESXi Shell running. Recommendation: switch off after administrative work.
- Lockdown mode
- Is off. Recommendation: lockdown mode (normal), so the host is managed through vCenter only.
- Time
- No NTP server configured or the NTP service is not running. A wrong clock breaks sign-in, logs and certificates.
- Logs
- No syslog target (Syslog.global.logHost). Without a central syslog server, the logs do not survive a reboot.
- Shell timeout
- UserVars.ESXiShellInteractiveTimeOut is set to 0. Recommendation: a timeout, for example 900 seconds.
- VMs
- Virtual hardware older than vmx-14, emulated network adapters (E1000, E1000E) instead of vmxnet3, connected CD/DVD drives. Plus VMware Tools that are missing, not running or outdated.
- 01
Create a user
A user in vCenter with the “Read-only” role, assigned at the top and propagated to child objects.
- 02
Start Spherifyer
Place the program on a Windows or Linux server and install it as a service.
- 03
Enter vCenter
Enter address and credentials in the browser, choose the interval.
- 04
Filter the warnings
After the first collection run, select the Compliance category and work through it item by item.
