Last updated:
When upgrading a Citrix environment to CVAD 2507 LTSR, you may encounter an easily overlooked obstacle.
The SQL Server is up to date.
SQL Server 2019 is in use.
Everything looks correct.
Yet the Citrix site upgrade fails because of the SQL requirements.
The cause may lie in the database itself.
Upgrading SQL Server does not automatically raise the compatibility level of existing databases.
For example, an environment might look like this:
SQL Server: 2019
Citrix Site DB: compatibility level 110
Citrix Monitoring DB: compatibility level 110
Citrix Logging DB: compatibility level 110
And 110 means:
👉 SQL Server 2012
SQL Server is living in 2019, while the Citrix database still thinks it is 2012. 😄
A quick check:
SELECT
name AS DatabaseName,
compatibility_level AS CompatibilityLevel
FROM sys.databases
ORDER BY name;
The key levels:
110 = SQL Server 2012
120 = SQL Server 2014
130 = SQL Server 2016
140 = SQL Server 2017
150 = SQL Server 2019
160 = SQL Server 2022
170 = SQL Server 2025
On SQL Server 2019, the compatibility level can be adjusted, for example, with:
ALTER DATABASE [CitrixSite]
SET COMPATIBILITY_LEVEL = 150;
GO
ALTER DATABASE [CitrixLogging]
SET COMPATIBILITY_LEVEL = 150;
GO
ALTER DATABASE [CitrixMonitoring]
SET COMPATIBILITY_LEVEL = 150;
GO
⚠️ Of course, do not run this blindly.
Beforehand:
✅ Check the Citrix/SQL support matrix
✅ Back up the databases
✅ Check the actual database names
✅ Use a maintenance window
✅ Then test the compatibility level and Citrix site
And another point from practice:
During a CVAD 2507 site upgrade, also check the SQL/upgrade account being used. Missing permissions for required changes inside the Citrix databases can also lead to interesting troubleshooting.
My new Citrix pre-upgrade check
Before a major CVAD upgrade, this small query now belongs alongside the Controller, VDA, StoreFront and SQL versions in my checks:
SELECT name, compatibility_level
FROM sys.databases;
Takes 10 seconds.
Can save you a lengthy troubleshooting session during a maintenance window.
Finding in the Wild. 🔎
Sometimes it is not the SQL Server that is too old.
It is the database that has forgotten which year we are in. 😄
