Even more USERSTORE_TOKENPERM

Reading this and this post reminded me of our very bad experience with the USERSTORE_TOKENPERM cache.

When we ran into our USERSTORE_TOKENPERM issue last year there was no real awareness of the problem yet (SQL Server 2005 x64 - Enterprise Edition Build 2153). It was so bad we had to revert to SQL Server 2000 because things got so slow that SQL Server started to become unresponsive. After we got everything back to the original server with SQL Server 2000 we had a couple of weeks of detective work in front of us. Having an issue with this much impact needs to be investigated and the root cause must be found before you can even think about migrating again. We had done all the steps needed before migrating like stress testing, functional jobs, system jobs, ... and yet we did not see the behavior it was showing in production.

We have spent quite some time on the phone with an Escalation Engineer from Microsoft trying to pinpoint the root cause of our problem. After we convinced the engineer that the root cause was not parallelism as he suggested we finally got some vital information that helped us understand what was happening. When he told us about the USERSTORE_TOKENPERM cache and its behavior we started running some tests and now knew what we were looking for. Sure enough the size of this cache just kept growing and the remove count certainly did not.

select
distinct cc.cache_address,
cc.name,
cc.type,
ch.clock_hand,
cc.single_pages_kb + cc.multi_pages_kb as total_kb, cc.single_pages_in_use_kb + cc.multi_pages_in_use_kb as total_in_use_kb,
cc.entries_count, cc.entries_in_use_count, ch.removed_all_rounds_count, ch.removed_last_round_count, rounds_count
from sys.dm_os_memory_cache_counters cc join sys.dm_os_memory_cache_clock_hands ch on (cc.cache_address =
ch.cache_address)
where cc.type = 'USERSTORE_TOKENPERM'
order by total_kb desc

Since it was announced that this problem would be fixed in Service Pack 2 we decided to create some test scripts that really put stress on this cache. We tried to come as close as possible to our production situation at that time which was sp_executesql for the CUD operations (UpdateBatch). Putting pressure on this cache was quite easy, have different query text so they get cached as different query plans and execute them twice since the first time it just gets inserted in the cache and the problem comes with the lookup action. We forced this behavior by executing queries with varchar parameters that increase in length every iteration. Also remember to use a login that is not a sysadmin since this causes SQL Server to skip the TokenAndPermUserStore check. We ended up creating a couple of scripts to do just this and also to monitor the cache size. The 'engine' behind it are simple command files and SQLCMD. Unfortunately I can not post the scripts since they are the property of the customer.

As far as we have seen from our tests the behavior has certainly improved in SP2. It takes a lot longer for the duration of the queries to increase and it does not increase to the point where SQL Server becomes unresponsive nor does the duration become unacceptably high. Since we are really pushing the cache to the limit it does still cause some increase in duration but not nearly to the point where SP1 gets it. We do notice however that under the constant 'hammering' of these queries even SP2 has the greatest trouble to keep the cache under control even when we call DBCC FREESYSTEMCACHE('TokenAndPermUserStore'). Obviously this scenario is not realistic since we really are trying to make it go wrong. Not only did the TokenAndPermUserStore behavior change in SP2 they cut down the size of the plan cache dramatically in SP2 too.

The conclusion however is that under Service Pack 2 we do not experience the same catastrophic behavior as before.

A lot of time has passed and we are going to migrate to SQL Server 2005 on the 9th of September. We are using a whole new .NET Framework on the application tier with a whole new data access layer, so a lot of the factors that existed that last time have now changed. We are starting our stress tests next month, if I get the chance and if we find anything peculiar I will let you know.

Hopefully this time there will not be any surprises :-(

The target of 'replace' must be at most one node

One of the new features of SQL Server 2005 is the integration of XML in the database. Although XML in the database is not my favorite architecture there are situations where it is useful. Since we have not really used XML in a real life project before we are slowly learning its limitations.

A limitation we recently discovered is modifying more than 1 node at a time using XML DML. This limitation is documented in the Books Online under the "replace value of" topic: "If multiple nodes are selected, an error is raised."

The only idea we could come up with is to loop through the nodes until all the nodes have been updated. Lucky for us SQL Server also supports mixing XML with special functions like sql:variable() and sql:column().

Check out the code here since blogger is giving me issues with the XML string :-(

Windows Home Server - RTM

There was not such a great buzz about this edition (at least not here in Belgium) but nevertheless it is an interesting product. It contains some nice features like a home network health monitor which makes sure all updates are applied on your home PC's, full system restores, automatic backup and many more.

Check out the announcement here and more information here

SQL Server 2008 Launch

SQL Server 2008, together with Visual Studio 2008 and Windows Server 2008, will be launched on the 27th of February. More information can be found here.

Greg Low once launched a nice suggestion for deferred constraints but unfortunately it did not make it to SQL Server 2008. But not to worry there are lot of other exciting features ;-)

Also some nice features in Windows 2008, for the I/O lovers: there is an internal limit of 64k per I/O request in earlier version of Windows, this limit has been removed in the new Windows core.

For Visual Studio 2008 we will see the integration of WF, WCF and WPF and of course LINQ.

Seeing the end of February date reminded me of a really stupid joke someone sent me.
So for the geeks among us:

It is March 1st and the first day of DBMS school
The teacher starts off with a role call.. Teacher: Oracle?
"Present sir"
Teacher: DB2?
"Present sir"
Teacher: SQL Server?
"Present sir"
Teacher: MySQL?
[Silence]
Teacher: MySQL?
[Silence]
Teacher: Where the hell is MySQL
[In rushes MySQL, unshaved, hair a mess]
Teacher: Where have you been MySQL
"Sorry sir I thought it was February 31st"

SQL Server 2005 Best Practices Analyzer (July 2007)

Although the summer (which we have not seen much of in Belgium the last 2 weeks) has started, Microsoft is still on a roll. Check out the new SQL Server 2005 Best Practices Analyzer here.

By the way, for those of you who have not bought SQL Server 2005 Practical Troubleshooting: The Database Engine, it is really a must read for every SQL Server DBA. Apart from troubleshooting information it also contains a lot of 'inside' information about how SQL Server 2005 works, I just love that stuff but maybe that's just me :-)

*EDIT*
A nice blog post here.

SQL DMVStats

Tom Davidson and Sanjay Mishra bring us a great set of reports and scripts to generate a Dynamic Management View Performance Data Warehouse.
Check out this great solution here.

This is actually something we had on our to do list for quite some time so this will save us a lot of development work.

Synonym

Perhaps a little known feature of SQL Server 2005 but it might come in handy when you are slowly changing your database to use schema's. Synonyms allow you to define an alias for your objects and this is very helpful when you are migrating to schema's.

By using synonyms you can change your object names while retaining your 'interface' to the application and this allows for a smoother and more phased migration.

A little example with a table but remember that this also works for stored procedures, functions, etc. and it is also possible to create synonyms for remote objects (linked server).


--Create a schema
CREATE SCHEMA SchemasRock
GO
--Create a table in the default schema
CREATE TABLE dbo.myTable
(ID int)
GO
--Insert a row
INSERT INTO dbo.myTable VALUES (1)
--Move the table to the new schema
ALTER SCHEMA SchemasRock TRANSFER dbo.myTable
--Insert another row (will fail!)
INSERT INTO dbo.myTable VALUES (2)
/*
Msg 208, Level 16, State 1, Line 1
Invalid object name 'dbo.myTable'.
*/
--Create a synonym to support 'old' table name
CREATE SYNONYM dbo.myTable FOR SchemasRock.myTable
--Insert another row
INSERT INTO dbo.myTable VALUES (2)
--Select the rows
SELECT * FROM dbo.myTable
SELECT * FROM SchemasRock.myTable
--Clean
DROP SYNONYM [dbo].[myTable]
DROP TABLE [SchemasRock].[myTable]
DROP SCHEMA [SchemasRock]

SQL Server 2005 SP2 Cumulative Hotfix 3175

We're at build 3175 with KB936305, the hotfix is not publicly available though.
For an overview of the hotfixes released after SP2 check out KB937137, it contains the list of fixes per build.

SQL Server 2005 Books Online (May 2007)

Back to the current release of SQL Server ;-)
Get the updated Books Online here.

Fun with the new MERGE statement

I found some time to play with CTP3 and checked out the MERGE statement. What you achieve with it may not be rocket science and was perfectly possible with a couple of IF statements but using MERGE gives you much cleaner code. It looks like Microsoft is investing a lot in 'cleaner' coding in SQL Server 2008, which I can only welcome with great enthusiasm.

A simple example will show you exactly what the MERGE statement does.

CREATE TABLE StagedData
(ID int, UserName varchar(20))
GO

CREATE TABLE RealData
(ID int, UserName varchar(20))
GO

INSERT INTO StagedData VALUES (1, 'Slava') , (2, 'Paul') , (3, 'Wesley')
GO

MERGE RealData r
USING
(SELECT ID, UserName FROM StagedData) staging
ON (r.ID = staging.ID)
WHEN MATCHED THEN UPDATE SET r.UserName = staging.UserName
WHEN TARGET NOT MATCHED THEN INSERT VALUES (ID, UserName)
WHEN SOURCE NOT MATCHED THEN DELETE;
GO
SELECT * FROM RealData
GO

UPDATE StagedData SET UserName = 'Kimberley' WHERE ID = 3
GO
MERGE RealData r
USING
(SELECT ID, UserName FROM StagedData) staging
ON (r.ID = staging.ID)
WHEN MATCHED THEN UPDATE SET r.UserName = staging.UserName
WHEN TARGET NOT MATCHED THEN INSERT VALUES (ID, UserName)
WHEN SOURCE NOT MATCHED THEN DELETE;
GO
SELECT * FROM RealData
GO

DELETE FROM StagedData WHERE ID = 3
GO
MERGE RealData r
USING
(SELECT ID, UserName FROM StagedData) staging
ON (r.ID = staging.ID)
WHEN MATCHED THEN UPDATE SET r.UserName = staging.UserName
WHEN TARGET NOT MATCHED THEN INSERT VALUES (ID, UserName)
WHEN SOURCE NOT MATCHED THEN DELETE;
GO
SELECT * FROM RealData
GO

DROP TABLE StagedData
GO
DROP TABLE RealData
GO

SQL Server 2008 - A first look

I took some time to look at the ''nifty" features in SQL Server 2008. There is a long list of new features shown in this webcast, but I doesn't look like all of them are already implemented in CTP3.

Here are some of the new features that are in this build and are quite nifty:

• MERGE, a very powerful statement to combine insert/updates and even deletes in a single statement
• Multiple values INSERT (eg. INSERT INTO myTable VALUES (1, 'SQL Server 2008 Rocks'), (2, 'It really does') - the example in the BOL seems to be incorrect)
• FORCESEEK query hint
• Table-valued parameters enables you to pass table variables to stored procedures or functions
• ALTER DATABASE SET COMPATIBILITY_LEVEL isof sp_dbcmptlevel
• Policy based management, this is an awesome feature that allows you to set policies on many configuration options and even naming conventions for your objects in a very flexible way (Declarative Management Framework)
• Change Data Capture, a built-in auditing mechanism
• Some new dynamic management views (sys.dm_os_memory_brokers, sys.dm_os_memory_nodes, sys.dm_os_nodes, sys.dm_os_process_memory and sys.dm_os_sys_memory)

There are probably many cool features and I'll keep you informed when I stumble upon them.

*EDIT*
A great list of new features can be found here.

Is it a bird, is it an airplane? No it's the first CTP of SQL Server Katmai... I mean SQL Server 2008!

Check it out here, the first public CTP of SQL Server 2008.

Go go go :-)

SQL Server Katmai Part II

Francois Ajenstat announced in a podcast interview - at the very first Microsoft Business Intelligence Conference - that the first CTP will probably be released in the coming month!

Get your Virtual Servers ready ;-)

By the way, you can find webcasts of this conference on the resources page.

3GB PAE AWE

I once posted some information about the 3GB, AWE and PAE settings.
Here is some more detailed information about these settings. Definitely worth reading!

Database Snapshot vs Procedure Cache

I recently stumbled upon a KB article (KB917828) that points out some situations in which your procedure cache is cleared.

Since one of our architectures for reporting involved database snapshots I was suprised to read that your whole procedure cache is cleared when you drop a database snapshot. This may not be as dramatic as it sounds but you should closely watch the impact on your environment when the cache is cleared.

Other situations are:

• A database has the AUTO_CLOSE database option set to ON. When no user connection references or uses the database, the background task tries to close and shut down the database automatically.
• A database has the AUTO_CLOSE database option set to ON. Maintenance operations are performed, such as the DBCC CHECKDB operation or a backup operation. When the operations finish, the background task tries to close and shut down the database automatically.
• You run several queries against a database that has default options. Then, the database is dropped.
• You change the database state to OFFLINE or ONLINE.
• You successfully rebuild the transaction log for a database.
• You restore a database backup.

SQL Server Katmai

It is announced to be available in 2008.
I'm ready for a beta/ctp version ;-)

More information here and here.

Datasheet here.

Visual Studio 2005 Team Edition for Database Professionals - SR1 (CTP)

For those of you who are lucky and are already working with VSDBPRO, there is a service release CTP available for download (KB936202). Make sure you read the release notes before you install.

Gert Drapers also posted some additional information on his blog.

One of the biggest enhancements, in my opinion, is the support for SQLCMD variables. This enables great flexibility in your deployment scripts.

Cumulative waitstats per session/query

We were troubleshooting a long running batch job when we suddenly realized that there is no easy way to get cumulative wait information for a specific session or query. With the rich information already available to us in SQL Server 2005 this might be a great addition to the product.

I decided to launch a suggestion on Microsoft Connect. This great initiative makes it possible for everyone to give suggestions to improve Microsoft's products.

Check out my suggestion here (and while you are there do vote ;-))

A couple of interesting KB articles

Today we were installing SQL Server 2005 on our new Windows 2003 Server (SP2) cluster when we ran into the following issue: "The file C:\Windows\Microsoft.NET\Framework\Meaningless_string\mscorlib.tlb could not be loaded", the solution to this can be found in KB918685.

Dirk G from MCS Belgium was nice enough to point us to another hotfix concerning the USERSTORE_TOKENPERM issue, KB933564 contains more information about this.

KB935897 has some information about the new "Incremental Servicing Model" for SQL Server.

So what's up with all the SP2 post fixes?

It's getting quite cumbersome to understand all the releases after SP2 for SQL Server 2005. Bob Ward from PSS decided to clarify this on his blog.

Aaron Bertrand also clarifies the differences between a GDR and a QFE here.