I’ve added a new script to my GitHub PowerShell repository for managing the IIS WMSvc Certificate. The script WMSvc_InstallCertificate.ps1 is intended to be ran either by scheduled task or by command line and will attempt to detect when the Web Management Service (WMSvc) certificate needs to be replaced. Replacement certificates are sourced from an Enterprise Certificate Authority automatically. The IIS machine account must have privileges to enroll.
In my lab, this script is tied to a scheduled task that is automatically created by group policy on servers attached to the IIS role security group. This same security group is also granted enroll privileges on the certificate template. In effect, new servers created will automatically receive a trusted certificate for their management port and that certificate is rotated automatically before expiration.
Since my lab IIS installations run on Server Core, it is quite convenient to have the remote management service configured automatically.
Starting with Windows 10 1511, Microsoft introduced a new feature called Secure Time Seeding, part of the Secure Time Service (STS), as an upgrade to the W32TIME service. The STS uses information from SSL connections to validate NTP data. Information from this feature supersedes all other time sources, including locally configured NTP, domain controllers, and Hyper-V time synchronization.
I first noticed the feature when several of my Hyper-V virtual machines began shifting their system clocks backwards and forwards several times a minute. At first the time changes spanned a few hours, but as the machine uptime climbed, so did the time jumps. Eventually, the time was bouncing backwards and forwards by weeks, several times a minute. The Hyper-V time synchronization service was fighting with the new Secure Time Service and this wrecked havoc on authentication and any other services running on the systems.
An issue exists between the Ubiquiti Networks Unifi software controller v5.10.19 and a Microsoft Radius or Network Policy Server (NPS) when NTLMv2 responses are forced on all Domain Controllers. When this condition exists, radius calls from the Unifi controller fail with error event that indicates the password is invalid. The security logs on the radius server report the following failure:
Failure Reason: Unknown user name or bad password.
Sub Status: 0xC000006A (Account logon with misspelled or bad password)
The issue is limited only to VPN connections. WiFi connections using the same Radius connection profile and user account succeed.
Applying the registry key in Microsoft KB 2811487 to the Radius server resolves the issue.
HKEY_LOCAL_MACHINE\System\CurrentControlSet\Services\RemoteAccess\Policy\Enable NTLMv2 Compatibility = 1
I spent a portion of the last two weekends attempting to upgrade my homelab vCenter Server running on Server 2016 from 6.5.0 to 6.5.0-U1 with little success. My implementation is only a month old and non-complex, aside from the external Platform Service Controller servicing the single vCenter Server (so far).
Both the PSC and the VCS returned the same error during installation: “Installation of component VCSServiceManager failed with error code ‘1603’. Check the logs for more details.”
I built a new computer over the weekend and wanted to take advantage of Storage Spaces to group some drives together. Little did I know that adding BitLocker to the volume would quickly corrupt and obliterate the data on it.
The issue only appears relevant on the Windows 10 1511 Novenber Build (TH2, 10586) when the OS drive is also encrypted by BitLocker and when the new XTS-AES 128 format is used. Shortly after encryption begins, files and directories will begin to vanish from the volume. After a reboot the volume is entirely unmountable and the above error displays stating that “the disk structure is corrupted and unreadable.”