Finding in the Wild | Citrix PVS – Database Connection Drops Because the vDisk Version Is 20!

Sometimes the smallest details lead to surprisingly extensive troubleshooting.

I recently received a report of particularly unusual behavior in Citrix Provisioning Services (PVS):

⚠️ The symptoms

The PVS environment experienced repeated, brief database connection interruptions.

The following events appeared regularly in the event log:

• Event IDs 201 / 214 – PVS database connectivity issues
• SQL Error 8023 – Error processing a supplied value
• PVS loses its database connection for a few seconds and temporarily goes offline.

The strange part: not all targets were affected, only one specific group.

🔬 The cause – reproduced by Citrix in the lab

After extensive analysis, Citrix reproduced the issue in its lab using the supplied data.

The cause lies in the Microsoft OLE DB Driver for SQL Server (MSOLEDBSQL), which PVS uses to communicate with its SQL database.

The decisive trigger?

vDisk version number 20!

When a PVS target starts or stops, PVS writes a license usage record to the database. This record also includes the version number of the vDisk in use.

If that version number is exactly 20, the driver transmits the value incorrectly.

SQL Server rejects the operation with Error 8023.

PVS interprets this error as a loss of database connectivity and briefly goes offline.

In the investigated environment, exactly one active vDisk was running version 20. This also explained why only the associated targets were affected.

🛠 The workaround

The solution is surprisingly straightforward:

✅ Do not use vDisk version 20 in production.
✅ Deploy a newer version, 21 or higher.
✅ Alternatively, merge the vDisk versions as appropriate.
✅ For other vDisks, avoid keeping version 20 in use long-term.

Important: this is not a general SQL connectivity issue. Restarting SQL Server or adjusting network timeouts does not eliminate the underlying trigger.

📌 Vendor status

Citrix has reproduced the behavior in its lab.

A permanent fix is planned for a future PVS release.

💡 My takeaway

A single version number can be enough to cause brief PVS database connectivity failures.

If you see PVS Events 201/214 together with SQL Error 8023, check the vDisk versions in use as well.

The number 20 certainly would not have been my first suspect. 😉

Finding in the Wild | When Identical Machine SIDs Suddenly Become a Citrix Problem

For years, this was completely normal in Citrix environments: Create a golden image → use MCS or PVS to provision hundreds of VDAs → done. Machines from the same image may share the same local machine SID .

For years, this was completely normal in Citrix environments:
Create a golden image → use MCS or PVS to provision hundreds of VDAs → done.

Machines from the same image may share the same local machine SID.
Usually, that was not a major problem.

After all, each computer has its own Active Directory computer account with a unique domain SID.

With stricter Windows security hardening, however, this old architectural assumption becomes interesting.

What can happen?
When two cloned VDAs communicate, certain machine-to-machine connections may be blocked because their local identities are identical.

For example, this may affect:
❌ RDP between VDAs
❌ SMB / file sharing
❌ Printer sharing
❌ Certain NTLM/Kerberos scenarios

The Windows event log may show LsaSrv Event ID 6167.

And this is where it gets interesting:
The VDAs are correctly domain joined.
The AD computer accounts are unique.
DNS is correct.
Kerberos also initially looks fine.
Yet the connection is blocked.

Why?
Microsoft is increasingly tightening machine identity validation.
A design decision that was unproblematic for cloned Windows systems for years suddenly collides with more modern Windows security hardening.

This does not affect an obscure Citrix feature.

It affects one of the basic principles of:
MCS and PVS: creating many machines from a single golden image.

Citrix itself is now emphasizing the security of local persistent accounts in shared MCS/PVS images and explicitly recommends not creating such accounts in the master image.

What does this mean in practice?
If communication suddenly fails after Windows security updates between two VDAs in the same catalog:
Do not immediately dismantle firewalls, DNS or Citrix policies. 😉

Also check:
System event log →
LsaSrv
and particularly Event ID 6167.

For me, this is a good example of why golden-image architectures now need testing against more than just new Citrix versions.

Finding in the Wild | PVS 2507 – When the Device License Suddenly Disappears

At a customer, I recently encountered an old acquaintance: “No device license is currently available for this computer. A system shutdown will be initiated in 720 hours. ” A classic PVS error message. With PVS 2507 + LAS

At a customer, I recently encountered an old acquaintance:
“No device license is currently available for this computer. A system shutdown will be initiated in 720 hours.”

A classic PVS error message.
With PVS 2507 + LAS, however, it is worth taking a closer look.

For example, the community has an interesting report involving PVS 2507 on Windows Server 2025:

Licensing initially works, but after approximately 23 hours PVS loses its licensing. Running the PVS Configuration Wizard again restores it. Until the following day.

A community finding? Yes.

But Citrix has now also officially fixed several interesting PVS licensing issues in 2507.1 CU2:
🔴 CVADHELP-31063
After migration to Citrix License Activation Service (LAS), PVS could incorrectly report Not entitled. Processing of expiration dates and time zones was among the factors involved.
🔴 CVADHELP-31165
Even time differences between PVS servers could cause device licensing issues. With CU2, PVS now tolerates up to 15 minutes of clock difference.

Another familiar issue after a LAS migration is :
Cloud vs. on-premises licensing
.

For CVAD on-premises + PVS + UHMC/Platform licensing, check that PVS is actually configured for on-premises use and that the corresponding PVS_CCS / PVSD_CCS entitlements exist.

My current PVS 2507 licensing checks therefore look like this:
✅ Check the exact PVS build
✅ Account for 2507.1 CU2 / known fixes
✅ Check LAS status and entitlements
✅ Check cloud vs. on-premises
✅ Check PVS_CCS / PVSD_CCS
✅ Compare NTP/clock settings across all PVS servers
✅ Consider the Stream Service after licensing changes

Only then would I start investigating the conventional route through ports, firewalls and the license server.

The interesting part
:
The error message is old.
Its possible causes are
not.

Anyone simply following the old CTX132198 checklist for PVS 2507 may easily overlook the newer LAS/2507-specific causes.

Finding in the Wild | PVS + Secure Boot – A Small CU2 Feature That Could Become Important in 2026

Citrix Provisioning 2507.1 LTSR CU2 is here. Among the usual fixes is a change that I find much more interesting for PVS environments than the version number suggests: 👉 Support for Microsoft Windows UEFI CA 2023 Why is

Citrix Provisioning 2507.1 LTSR CU2 is here.

Among the usual fixes is a change that I find much more interesting for PVS environments than the version number suggests:

👉 Support for Microsoft Windows UEFI CA 2023

Why is this important?

Microsoft is currently renewing the certificate chain behind UEFI Secure Boot. Some existing Secure Boot certificates date back to 2011, and Microsoft is gradually migrating Windows to the new 2023 certificates.

This is exactly where it becomes interesting for Citrix PVS.

PVS must understand the new Secure Boot world

With PVS 2507.1 LTSR CU2, Citrix now supports Secure Boot with images signed using the new Windows UEFI CA 2023 certificate.

Initially, this sounds like a small compatibility fix.

But it is not necessarily.

Anyone familiar with PVS knows:

💥 If the boot chain is wrong, there is no Citrix session.

We are no longer discussing logon duration, profiles or HDX performance.

The target simply does not boot.

Another interesting change

When a target device starts, PVS now automatically detects whether the system booted using:

BIOS or UEFI

.

This information is saved in the device record and is also available through farm telemetry.

I find this particularly useful in larger or long-established PVS farms.

Different firmware and boot configurations are surprisingly common there.

What would I check now?

Anyone using PVS + Windows 11 or Windows Server 2025 + Secure Boot should definitely keep CU2 on their radar.

Before a rollout, I would test in particular:

👉 Secure Boot + Windows UEFI CA 2023
👉 Target device boot
👉 PXE / BDM
👉 Existing vDisks
👉 Hypervisor firmware / UEFI configuration
👉 Boot after promoting a vDisk version
👉 Failover to a second PVS server

With changes like these:

The vDisk has worked for years is no guarantee that the future Secure Boot chain will also work.

My takeaway

For me, PVS 2507.1 CU2 is therefore more than just:

Another cumulative update with a few bug fixes.

Support for Windows UEFI CA 2023 is a classic future-readiness topic.

Microsoft is changing the Secure Boot trust chain.

Citrix is following suit with PVS.

And once a target fails to start after a Windows/Secure Boot update, an inconspicuous release-note entry suddenly becomes an important feature. 😎

Finding in the Wild | PVS + XenServer – Cores per Socket

A small detail that can lead to extended troubleshooting with Citrix PVS on XenServer. When provisioning VMs through the CVAD Setup Wizard, the number of vCPUs must match the CPU topology of the XenServer template. The vCPU count

A small detail that can lead to extended troubleshooting with Citrix PVS on XenServer.

When provisioning VMs through the CVAD Setup Wizard, the number of vCPUs must match the CPU topology of the XenServer template.

The vCPU count must be a multiple of Cores per Socket.

Example:

Cores per Socket: 4
PVS Wizard: 6 vCPUs ❌
PVS Wizard: 8 vCPUs ✅

With an invalid combination, the VM is provisioned but cannot subsequently boot.

XenServer may report:

VCPU_max must be a multiple of platforms:cores-per-socket

The tricky part:

The VM exists.
The PVS target exists.
The vDisk assignment looks correct.

Troubleshooting can quickly focus on PXE, BDM, streaming or the vDisk, although the actual issue is the VM’s CPU topology.

Citrix continues to track this as PVS-1126.

👉 So, for PVS targets that do not boot on XenServer, always check:

vCPU ↔ Cores per Socket.

A small setting with a big impact. 😉