First we had sp2 (build 3042)...
Then came sp2a (still build 3042 but 3043 if you looked at the dlls)
A hotfix was then released to bring those on old sp2 up to sp2a (which took the build number to 3050)...
Now we have the post sp2 cumulative hotfix release taking us to a whopping build no of 3152!
Details (and downloads) of which can be found here
In short it would appear that a regression has taken place between builds 2214 and 2219 and these fixes didn't make it to sp2.
I have cut and pasted the hotfix list for this rollup patch for your convenience below
Bug No Description
| 50000783 |
When you define the value of the AdminTimeout property and the value of the ServerTimeout property, a profiler trace stops unexpectedly. |
| 50000784 |
When you make a data manipulation language (DML) change to a published table, error 2812 occurs. |
| 50000785 |
When you create a cursor by using an SQL statement that uses the OPTION (RECOMPILE) hint, error 16943 occurs. Error 16943 indicates that the table schema changed. This problem occurs when you retrieve information from the cursor location. |
| 50000786 |
A memory leak exists in the MSOLAP provider (Msolap90.dll). |
| 50000787 |
When you configure the Scheduler component to use a time-out value that is less than 15 milliseconds, you experience high CPU usage. |
| 50000809 |
The LazyWriter process uses lots of CPU resources. This problem occurs on a non-uniform memory access (NUMA)-enabled x64-based server that is running SQL Server 2005 x64 Edition. |
| 50000810 |
When you update a view through a cursor, you cannot make a CSEQUAL request or a TSEQUAL request for rows that are extended by using NULL values. |
| 50000873 |
When you run an application in SQL Server 2005, error 8624 occurs. Error 8624 indicates that the cycle check failed. |
| 50000874 |
When the stored procedure runs outside an explicit transaction, a temporary table in a stored procedure is not cleaned up correctly. |
| 50000889 |
When you run a Microsoft Office Excel data mining operation, you receive the following error message:
The 'MINIMUM_DEPENDENCY_PROBABILITY' data mining parameter is not valid for the 'Table1_283833_NB_413946' model |
| 50000939 |
The Maintenance Cleanup task uses an incorrect age group for cleanup operations. |
This is disappointing news indeed and I fear not the end of the story. We'll have to just wait and see what else comes out.
Many thanks to Darren for the heads up!