This article describes the following about this hotfix release:

•	The issues that are fixed by this hotfix package
•	The prerequisites for applying the hotfix package
•	Whether you must restart the computer after you apply the hotfix package
•	Whether the hotfix package is replaced by any other hotfix package
•	Whether you must make any registry changes after you apply the hotfix package
•	The files that are contained in the hotfix package

Back to the top

When a custom application that is running on Microsoft SQL Server 2005 uses features that trigger frequent database protection timestamp changes, a gradual increase in memory consumption for the USERSTORE_TOKENPERM cache store occurs. Additionally, many duplicate TokenAccessResult entries have a class of 65535 in the sys.dm_os_memory_cache_entries dynamic management view.

For more information about the issue and about conditions that cause protection timestamp changes for a database, see the "More information" section.

Back to the top

This problem occurs because the cumulative permission check of a query is stored in the USERSTORE_TOKENPERM cache store as a TokenAccessResult entry that has a class of 65535. TokenAccessResult entries use the protection timestamp to determine whether security changes have occurred that would invalidate the cache entry. Every time that the protection timestamp changes, the old cache entries cannot be reused because the old entries may not be current. Therefore, a new cache entry must be inserted. However, an old entry is not removed until SQL Server experiences memory pressure. This problem can lead to an increase in memory consumption by the USERSTORE_TOKENPERM cache store.

Back to the top

A supported hotfix is available from Microsoft. However, this hotfix is intended to correct only the problem that is described in this article. Apply this hotfix only to systems that are experiencing this specific problem. This hotfix might receive additional testing. Therefore, if you are not severely affected by this problem, we recommend that you wait for the next software update that contains this hotfix.

If the hotfix is available for download, there is a "Hotfix download available" section at the top of this Knowledge Base article. If this section does not appear, submit a request to Microsoft Online Customer Services to obtain the hotfix. To submit an online request to obtain the hotfix, visit the following Microsoft Web site: Note If additional issues occur or if any troubleshooting is required, you might have to create a separate service request. The usual support costs will apply to additional support questions and issues that do not qualify for this specific hotfix. To create a separate service request, visit the following Microsoft Web site: Note The "Hotfix download available" section and the online request forms display the languages for which the hotfix is available. If you do not see your language, it is because a hotfix is not available for that language.

Prerequisites

You must have Microsoft SQL Server 2005 Service Pack 2 (SP2) installed to apply this hotfix.

For more information about how to obtain SQL Server 2005 Service Pack 2, click the following article number to view the article in the Microsoft Knowledge Base:

Restart information

You do not have to restart the computer after you apply this hotfix.

Registry information

You do not have to change the registry.

Hotfix file information

This hotfix contains only those files that are required to correct the issues that this article lists. This hotfix may not contain of all the files that you must have to fully update a product to the latest build.

The English version of this hotfix has the file attributes (or later file attributes) that are listed in the following table. The dates and times for these files are listed in Coordinated Universal Time (UTC). When you view the file information, it is converted to local time. To find the difference between UTC and local time, use the Time Zone tab in the Date and Time item in Control Panel.

SQL Server 2005, 32-bit versions

File name	File version	File size	Date	Time	Platform
Microsoft.sqlserver.maintenanceplantasks.dll	9.0.3153.0	296,304	08-Mar-2007	00:38	x86
Msmdlocal.dll	9.0.3153.0	15,930,224	08-Mar-2007	00:38	x86
Sqlaccess.dll	2005.90.3153.0	350,576	08-Mar-2007	00:38	x86
Sqlservr.exe	2005.90.3153.0	29,190,512	08-Mar-2007	00:38	x86

SQL Server 2005, 64-bit versions

File name	File version	File size	Date	Time	Platform
Microsoft.sqlserver.maintenanceplantasks.dll	9.0.3153.0	296,304	08-Mar-2007	00:38	x86
Msmdlocal.dll	9.0.3153.0	15,930,224	08-Mar-2007	00:38	x86
Sqlaccess.dll	2005.90.3153.0	357,744	08-Mar-2007	10:53	x86
Sqlservr.exe	2005.90.3153.0	38,638,960	08-Mar-2007	10:53	x64

Back to the top

Microsoft has confirmed that this is a problem in the Microsoft products that are listed in the "Applies to" section.

Back to the top

You can use the following two queries to determine whether you are experiencing this issue.

•

When you run the following query, you notice that memory consumption by the USERSTORE_TOKENPERM cache increases: select sum(single_pages_kb+multi_pages_kb) 'total memory for tokeperm' from sys.dm_os_memory_clerks where type = 'USERSTORE_TOKENPERM'

•

When you run the following query, you receive many entries in the Store Address column and in the ID column: select [Store Address], [id], count (*) 'number of entries' from (select cast(entry_data as xml).value ('(//@store_address)[1]', 'varchar (100)') as [Store Address], cast(entry_data as xml).value ('(//@id)[1]', 'bigint') as [id] from sys.dm_os_memory_cache_entries where type = 'USERSTORE_TOKENPERM' and cast(entry_data as xml).value ('(//@name)[1]', 'varchar (100)') = 'TokenAccessResult' and cast(entry_data as xml).value('(//@class)[1]', 'bigint') = 65535 ) R group by [Store Address], [id] having count (*) > 1 order by count (*) desc If this query produces no results, you are not experiencing the issue that this article describes.

Many conditions change the database protection timestamp. For example, most Data Definition Language (DDL) operations change the protection timestamp.

The Create Table DDL operation and the Drop Table DDL operation for temporary tables changes the protection timestamp in the tempdb database.

The following DDL operations are Transact-SQL statements. These operations change the protection timestamp in the master database:

•	Create Login
•	Alter Login
•	Drop Login
•	Create Endpoint
•	Alter Endpoint
•	Drop Endpoint

The following DDL operations are related to user objects, such as the table object. These operations change the protection timestamp in any database. These databases include the master database and the tempdb database.

•	Create
•	Alter
•	Drop

The following DDL operations are related to security. All security-related operations change the protection timestamp in any database. These databases include the master database and the tempdb database. The following list names some examples of security-related DDL operations.

•	Create
•	Alter
•	Drop user
•	Role
•	Application role
•	Certificate
•	Schema
•	Symmetric keys
•	Asymmetric keys

Additionally, operations that grant, revoke, or deny permissions on an object are related to security. These operations also change the protection timestamp in any database. These databases include the master database and the tempdb database.

Other conditions can also cause the USERSTORE_TOKENPERM cache to grow over time. The fix that this article describes is for a very specific condition. That is, a change in the protection timestamp causes the cache store to grow. For more information about the USERSTORE_TOKENPERM cache, click the following article number to view the article in the Microsoft Knowledge Base: After you install this hotfix, a new attribute that is named timestamp is added to entry_data columns in the sys.dm_os_memory_cache_entries view. This attribute specifies the number of times that SQL Server checks the permissions on each plan. When a plan is newly compiled or is recompiled, the timestamp is 1. This value is recalculated if the protection timestamp changes. You can use the following query to take snapshots of the timestamp:

select [store_address], [timestamp], count (*) 'number of entries'  from  
	(select 
 cast(entry_data as xml).value ('(//@store_address)[1]', 'varchar(100)')  as store_address ,
case cast(entry_data as xml).value ('(//@timestamp)[1]', 'int') when  1 then 1 
		else  null 
		end as [timestamp]
		 from sys.dm_os_memory_cache_entries
		where type = 'USERSTORE_TOKENPERM' and cast(entry_data as xml).value ('(//@name)[1]', 'varchar (100)') = 'TokenAccessResult' and 
			cast(entry_data as xml).value('(//@class)[1]', 'bigint') = 65535
	) R 
group by [store_address], [timestamp]
order by count (*) desc

If over time you see many timestamps of 1 for each store_address column, an application has a high rate of ad hoc queries or of recompile operations. You must slow down the rate of the ad hoc queries or of the recompile operations. Alternatively, you can address the issue by applying trace flag 4618 to limit the number of entries per user cache store. Using trace flag 4618 can incur a small CPU overhead because this trace flag removes old cache entries as new entries are inserted. The trace flag performs this action to limit the size of the cache store growth. However, the CPU overhead is spread over time.

Back to the top

For more information, click the following article number to view the article in the Microsoft Knowledge Base:

Back to the top