Fixed issues

MCS VDAs using MCSIO and Citrix Provisioning VDAs sometimes become unregistered due to hangs processing disk requests from windows when Asynchronous IO is enabled in MCSIO or the PVS vDisk. The fix forces synchronous handling of some requests even when Asynchronous IO is enabled.

Under some circumstances, the automated updates task would fail when dealing with targets on VMware because the VM that exists can't be found to finish the provisioning due to issues in how the code was caching information.

Under some circumstances, Windows would return a non-fatal error to Citrix Provisioning streaming service which would cause the specific thread to exit. Eventually all processing threads would exit at which point no target can log on to the server. The fix treats this error as a warning instead of a fatal error.

When running on Intel processors from 2020 or newer, some hypervisors injects delays in processing whenever a VM runs an unaligned interlocked instruction. The performance counters in the Citrix Provisioning target device drivers would sometimes do this due to incorrect alignment definitions.

When communicating with the Citrix License Server to upload usage telemetry, the connection would fail if the system clocks on the Citrix Provisioning and License Server were not synchronized exactly. This fix allows a slight clock drift between these servers. Also, you must synchronize the servers to the same time source to avoid future problems with clock drift.

The vDisk status icon on the Windows tray on target devices would sometimes disappear. This is because it would not be re-displayed if the explorer process that implements the desktop restarted.

When the recent change was made to add the ability to define read only admin groups, the powershell SDK had a bug whereby it was never possible to specify that read only access was required when grating a group admin access.

If running in a timezone east of GMT, when the Citrix Provisioning Server loses contact with the license server, it would incorrectly decrease the expiration date of the LAS activation by two days every day as it sometimes treated days as local time instead of UTC (so midnight local time is the previous day in UTC). The fix treats all license dates as UTC to avoid this problem.

If the clocks on Citrix Provisioning Servers in a farm are not synchronized exactly, the Citrix Provisioning Server where the clock is set to an earlier time than the server which checks license activation with the License Server would incorrectly declare that Citrix Provisioning Server is not licensed. The fix allows the clocks to drift up to 15 minutes when checking the cached data. Also, you must synchronize the servers to the same time source to avoid future problems with clock drift.

When uninstalling the Citrix Provisioning Server installation, a command window pops up running the Windows service control manager command.

The Configuration Wizard fails if you try to create a new site while a database key rotation is in progress. The fix blocks adding a server to the farm when you select to create a new site instead of letting the process run and fail at the very end.

Selecting Get Template Properties causes the Citrix Provisioning console to crash.

No audit record is created when selecting to revert a version of a vDisk.

If you attempt to perform a reverse image on a target VM with no local disk the error is misleading and incorrect. The error message now correctly mentions that there is no local disk available.

When streaming with IPv6, Citrix Provisioning packets might be fragmented that leads to boot issues.

When initializing database encryption, the code can mistakenly think a database encryption key rotation is in progress and report an error.

If the user running the Configuration Wizard is not a domain administrator, then an error dialog about SPN creation is popped up with just the permissions error. The fix changed the error to a warning pop-up (since it does not cause the Configuration Wizard to fail) along with pointers describing what can go wrong and how to fix it.

Failure to create SPN means that if you are running with NTLM disabled, the Citrix Provisioning console and Configuration Wizard running on other servers will not be able to communicate with the server being configured.

If a server becomes unresponsive while you handle storage read and write requests that require guaranteed persistence (Force Unit Access), VDAs might become unregistered.

In VMware vSphere environments, scheduled vDisk automated update tasks might fail with the error "VM is not available to process updates" because of stale datacenter inventory; PVS now refreshes the datacenter information and retries locating the virtual machine.

If transient network errors occur, such as ICMP TTL-exceeded messages, stream process threads might terminate and not restart, which can eventually prevent target devices from logging on to or connecting to the PVS server on the affected port; these errors are now logged as warnings and streaming continues.

If Windows taskbar (Explorer) restarts, such as after a crash or after you log off and log on, the vDisk status icon in the system tray might disappear permanently.

If you use the IsReadOnly parameter with the New-PvsAuthGroup or Set-PvsAuthGroup cmdlet, a database error might occur; use the Grant-PvsAuthGroup cmdlet to set the read-only state of an authorization group instead.

If the Maximum devices booting pacing threshold is reached, UEFI targets that are reconnecting might be blocked; booting devices are now told to pause and retry while reconnecting devices can still connect.

If PVS servers are in time zones east of GMT, Licensing might incorrectly display "Not entitled" after an LAS migration because license expiration dates shift to the previous day.

If minor clock differences exist between PVS servers, license entitlement checks might fail and target devices might display the message:

"A device license has been reacquired and the pending shutdown canceled."

At desktop sign-in; server clocks can now differ by up to 15 minutes without triggering the error.

If you repair or uninstall the PVS server, a command prompt window might briefly appear.

After you install the target device software, the copyright symbol might display incorrectly in the Imaging Wizard's Japanese and Simplified Chinese localized files.

If encryption key rotation is in progress, the Configuration Wizard might fail when you add a new site to an existing farm; it now displays a clear error message instead.

If you uninstall the PVS console, the Microsoft.Kiota.Abstractions.xml file might remain on the system.

In Azure and GCP environments, upgrading the PVS Server and Console from 2507.1 LTSR to 2507 LTSR CU1 might fail.

If you upgrade the PVS Server from 2507 LTSR CU1 to 2511, the upgrade might fail with Event Error 11920 because the Citrix Infra Monitor Service does not start.

If you test Azure host connection credentials, you might receive an "Object reference not set to an instance of an object" error because the required machine creation plugins are not properly registered during installation on Azure and GCP deployments.

If you select Get Template Properties while creating a device, an "Application Error" dialog might appear.

If you use the Revert version option for a vDisk, audit trail entries might not be recorded correctly.

If you reverse image to a virtual machine that has no local hard disk, you might see a misleading error instead of the message "A local disk was not found." A new P2PvsGpt command is also available for reverse imaging UEFI/GPT systems.

If PVS Server Identity Key rotation occurs, target devices might fail to authenticate after the key changes.

If you run the undojoin.sql database script, it might fail because of a syntax error.

If you configure SSL for the SOAP and REST services, PvsRpc might fail because the service cannot access the server certificate's private key; private key permissions are now automatically updated so the Network Service account has read access.

If you upgrade from 2402 CU3.1 to a newer release, the upgrade might fail during the "configuring services" step of the Configuration Wizard.

If you send PvsRpc messages to the inventory service, spurious crashes might occur.

In the Configuration Wizard, Japanese text might display incorrectly because of an encoding issue.

If you upgrade from an older PVS version, such as 2203 CU7.1, the Configuration Wizard might crash when no server public key is present in the database.

If you delete the corresponding device collection in PVS, Azure and GCP virtual machines provisioned to a DaaS Delivery Controller might not be deleted correctly.

If the Workspace Environment Management Agent prerequisite requests a restart during an upgrade, the PVS server installer might exit prematurely without errors.

If the database is offline for images that use the native PVS RPC transport, provisioning servers might crash.