My torn data pages - Saturday, July 24, 2010

Sometimes one is confronted with forgetting an option... It all started with a database that had a single data file of 1.8 GB and a transaction log more than 6 times the size of that. Usually that would be a database in full recovery model and no transaction log backups. This database was in simple recovery model though and according to the Disk Usage report, 99.99% of the transaction log was free. So a DBCC SHRINKFILE should solve the problem, however DBCC SHRINKFILE solved nothing, the log remained at about 11 GB.

A query on sys.databases learned that this particular database had ACTIVE_TRANSACTION for log_reuse_wait_desc. A peek in Activity Monitor then showed that there was an open transaction for the database; a DELETE task with a huge wait time and the BROKER_RECEIVE_WAITFOR wait type.

Then it occurred to me this is the database that used to have a deadlock problem. After implementing some procedure fixes, there haven’t been any deadlocks for over a half year and neither has the server (or service) been shutdown in between.

The stored procedure on the queue does a receive, but no timeout associated to it. Effectively this means that the receive will wait inside the transaction, until a messages comes into the queue. That hasn’t happened in the last six months. The resolution is simple, add the TIMEOUT option to the RECEIVE and the procedure will stop when there wasn’t any work to be done for the specified amount of time (in milliseconds). For the code, see the post “Deadlock Alerts Through Event Notification”, which has been updated with the TIMEOUT option.

After altering the procedure, kill the spid. Service Broker will restart, but now with the TIMEOUT option in the procedure and stop after the specified amount of time has elapsed without messages arriving in the queue.

When restoring a backup of a database, SQL Server recreates the files with the same sizes as listed in the backup, or at least tries to do so. The only option you have, is specifying the path for each file. Generally not a problem, but every now and then I encounter a database with a transaction log about 10 to 50 times the size of the data file(s) . . . yep, full recovery model and just daily full backups. I am the first to acknowledge transaction logs backups are key to prevent this situation, but the next best thing still is that recent valid database backup. . . and then you can’t restore it, because the 200 GB to write that big empty file called transaction log cannot be allocated.

Today was one of those days;

Msg 3257, Level 16, State 1, Line 1
There is insufficient free space on disk volume 'D:\' to create the database. The database requires 101197938688 additional free bytes, while only 19386400768 bytes are available.
Msg 3119, Level 16, State 4, Line 1
Problems were identified while planning for the RESTORE statement. Previous messages provide details.
Msg 3013, Level 16, State 1, Line 1
RESTORE DATABASE is terminating abnormally.

Surely I wouldn’t be the only one having problems with this and I found two others on connect;

• Option to resize datafile/transaction log during a backup or restore operation
• Restore option to restore minimum file sizes

I voted (as important) for both items, though the second one is already closed. Yet I was triggered by the comments from the MS-people to ask for a specific solution towards the transaction log;

RESTORE DATABASE ... WITH STUB_LOG

One of the habits of RESTORE DATABASE is that it wants to allocate the log size (and files) as found in the backup. Strange, as the size of those files is the result of past activity (not contained in the backup being restored). The only space needed in the transaction log, is that of the activity during the backup. Additional space requirements may exist for subsequent restores of differential backups and log backups (in norecovery or standby) or future activity (after recovery), but that isn’t an issue as the log files would grow again during these restores.

The only real issue I can think of is having a database with fixed size transaction log file(s), which means no auto grow. So for a WITH STUB_LOG option to work properly in these scenarios, it may need to imply that the log file(s) are always set to unrestricted auto grow (possibly only honoring the original increment).

It would at least prevent me from having to allocate 100GB for an empty transaction log on 20GB available.

If you're interested in expanding your skills in SQL Server, check out these events;

• Quest Virtual Training for SQL Server, an online and worldwide available event on Wednesday March 3^rd
• SQL Zaterdag, hosted at InfoSupport in Veenendaal, Netherlands at Saturday March 6^th

Happy learning.

Microsoft recently launched a Virtual Business Card site for Microsoft Certified Professionals. So if you ever passed one or more of those Microsoft exams in the NT 4.0 or later eras, check out the www.mcpvirtualbusinesscard.com site (Windows Live ID sign in to connect to you records in Microsoft's certification database).

I set my profile page up to show off the credentials..

Also had a look at including the transcript, but that would be a 13-page pdf, a bit of overkill if you ask me.

Yesterday I visited the SQL Server day 2009, organized by the Belgian SQL Server User Group SQLUG.BE, in Mechelen. Congratulations on the event guys!

After the keynote by Microsoft Belgium (I wish they had talked a little bit more about SQL Azure), I visited the session by Henk van der Valk on world record ETL... now Henk has control over some top notch iron, but that doesn't mean his tips don't apply to a modest setup. Henk also mentioned he recently joined the blogsphere at www.henkvandervalk.com.

Next I sat (sorry I have to say so) horrible sponsored session by Quest... and this has nothing to do with the FogLight product. On another occasion (an afternoon session by Quest Netherlands) I witnessed the possibilities of Foglight (for .NET, SQL Server and VMware) and I must say it's a good looking product. However we got 30 minutes of boring listing of challenges and day to day problems (as if we weren't aware of them already) and in the end got some screenshots, which were completely out of context. I would have be completely lost in the presentation if I hadn't been to the session earlier by Quest NL.

After that, I meant to sit the session "Reporting Services a DBA's tool? YES!!!", but since the agenda card was a little confusing (or better said, I didn't pay enough attention) I walked into the session by Nico Jacobs on "Handling XML in SQL Server". Funny, as there was nothing new for me in the session but still I really enjoyed it... most important because as a trainer you rarely get an opportunity to see one of your colleagues at work on a familiar subject. Thanks Nico, I really enjoyed it.

The other session I attended was on "Policy Based Management", again by Nico Jacobs. I hoped it would go deeper into the possibilities of implementing it for SQL Server 2005 and SQL Server 2000. Unfortunately that was not the case, so I'll have to dive into the Enterprise Policy Management Framework and PowerShell without a quick start. But again, it was a joy listening to Nico.

Final session and closing keynote was by Chris Webb on PowerPivot (a.k.a. Gemini). It wasn't my first glance at Gemini, but it definitely is the kind of quick start I was looking for. Sitting a session like that saves a day of looking for stuff.

All-in-all, a day well spent.

Creating linked servers between two instances of SQL Server may not always be as straight forward as hoped. To make it more complicated, interference may be caused by matters beyond the control of the database administrator; Active Directory. After creating a linked server, being able to successfully use it may be the real challenge, it may even work for one user and fail for another.

Service Account

To communicate successfully in a domain environment, the service account for your database engine is the first to check/configure. Possible choices for the service account are;

• Domain User (recommended)
• Local User
• LocalSystem
• LocalService
• NetworkService (recommended)

From those options, a local user or LocalService for the service account will eventually stop you. In both cases a SQL Server has the permission set of user on the machine where SQL Server operates, with no identity on the network. In case the SQL Server has to connect to any remote resource, it will do so as an anonymous user (null session). Unless you want to be absolutely sure that SQL Server should never do anything beyond anonymous on the network, you can go this route... but hey, you're reading about linked servers, so just drop the thought of using a local user or LocalService.

Valid options to enable SQL Server to operate on the network are LocalSystem, NetworkService and Domain User. All have the potential to work well, but LocalSystem is not recommended from a security perspective. LocalSystem has the permission set of an administrator on the system where SQL Server operates, anyone obtaining control over your SQL Server has full access to the underlying operating system (think about the damage that could be done when xp_cmdshell is enabled). From a network perspective, LocalSystem and NetworkService are identical; when the SQL Server operates over the network, it will do so under the credentials of the Windows machine. With a domain user as service account, SQL Server has user permissions locally (as is the case with NetworkService, LocalService and a local user), on the network SQL Server will present the credentials of the domain user.

So two good options remain; NetworkService and a domain user. Which one you chose mainly depends on the advantages of one over the other;

• NetworkService:
  • No account or password management in Active Directory
  • Service Principal Names are automatically generated
• Domain User
  • Multiple instances on different machines can have the same identity which is required for clustering.
  • Transparent security for Database Mirroring. 
  • Usually SQL Agent will be the same account, this makes things easier with Log Shipping and Replication.
  • Improved access to Service Master Key (in case you don't have a backup of the key itself). -->

Personally, I favor using a domain user.

Should you find that you need to change the Service Account; use the "SQL Server Configuration Manager"

Configuring and using a SQL Server linked server for Windows Authentication

There are two ways to go about the configuration; graphically and via stored procedure. The graphical approach brings up a dialog, where just a little configuration is needed.

On the General page, select the server type to be SQL Server (choosing SQL Server Native or OLEDB drivers will work too, but require extra configuration) and type the name of the target instance using the flat name or FQDN followed by a backslash and instance name in case of a named instance.

On the Security page, select the option to "Be made using the login's current security context", which effectively tells SQL Server to impersonate or delegate the login requesting access though the linked server.

On the Server Options page, verify that Data Access is set to True. To see the stored procedures involved, either use the Script button on top of the dialog instead of pressing ok or right-click a configured linked server and "Script Linked Server as >".

To avoid running into (later explained) problems when creating linked servers, create the linked server from a shell (interactive or remote desktop) on the Windows machine where the linked server is configured.

To use the linked server, issue the query

SELECT TOP 50 * FROM [HOST.NET.LOCAL\INSTANCE].AdventureWorks.Sales.SalesOrderDetail

Note that the linked server HOST.NET.LOCAL\INSTANCE is enclosed in braces, which is required when the server name is fully qualified and/or when it is a named instance. If you don't want to expose host and/or instance names, consider creating an alias in "SQL Server Configuration Manager" under client configuration and point the linked server to the alias.

Impersonation and delegation

During the configuration (Security page) impersonation and delegation were mentioned. Also to avoid problems while configuring a linked server, the best way to do so was on the server where the linked server was to be created. The reason to do so is to avoid delegation. Below are a summary of impersonation and delegation (and links to a more complete overview). Windows accounts will be impersonated if necessary, Windows accounts will not be delegated unless both account and delegating system are configured to do so.

Impersonation happens when a process connects to resources pretending to have the identity of it's caller. Like when a user Bobbie is connected to SQLINST1 and issues a query for a table via a linked server on SQLINST2. Depending on the configuration of the linked server, SQLINST1 will pretend to be Bobbie when connecting on Bobbies behalf to SQLINST2. For a more detailed overview of impersonation, check the Pluralsight-wiki on impersonation. Note that impersonation is limited to the  system where process making the call is active.

Delegation is much like impersonation, with the additional feature that delegation is not limited to the system of the process making the call. Because the process now has to forward the credentials to another system and prove to operate on behalf of someone else than the system itself, the Kerberos protocol is required (that or knowledge of the password of the delegated user). For a more detailed overview of delegation, check the Pluralsight-wiki on delegation.

From the perspective of Windows systems, Bobbie (DOMAIN\Bobbie) is logged in on a typical client system (Windows XP, Vista or 7) and uses an application that connects to SERVER1\SQLINST1. In SQLINST1 two linked servers are configured; SERVER1 (being the default instance on SERVER1) and SERVER2\SQLINST1. The context of the application is DOMAIN\Bobbie and thus an access token for DOMAIN\Bobbie is handed to SERVER1 (for logging in on the SQL Server service SERVER1\SQLINST1). When the application issues a query for an object on the default instance of SERVER1, one local process on SERVER1 has to impersonate DOMAIN\Bobbie to the other local process on SERVER1. This will succeed and can work both with NTLM and Kerberos, because SERVER1 trusts itself and the domain; DOMAIN\Bobbie was authenticated by the domain and is only used by SERVER1 locally.

Next the application calls for SERVER2 and now the challenge is on! Bobbie get's a

Msg 18456, Level 14, State 1, Line 1
Login failed for user 'NT AUTHORITY\ANONYMOUS LOGON'.

SERVER2 does not trust SERVER1 and therefor will not accept the claim of SERVER1 that it acts on Bobbies behalf, SERVER2 wants to see PROOF! One way of proof would be if SERVER1 had the password of Bobbie, indicating that Bobbie fully trusts SERVER1... but Lesson number 1; never give anyone your password. Lesson number 2; don't write it on a Post-It... Lesson number 3; don't trust computers that violate lessons 1 or 2. The other option is to turn to a mutually trusted third party: the domain, but now the configuration needs to be right for Kerberos.

For Kerberos to do it's trick, the user (Domain\Bobbie) and the machine delegating (SERVER1) the user have to be trusted for delegation. This is sufficient if the SQL Server services on SERVER1\SQLINST1 and SERVER2 run as NetworkService.

In case SERVER1\SQLINST1 is running under a domain user account, trust that user account for delegation, not the server.

Service Principal Name

Kerberos distinguishes three roles;

• Key Distribution Center (KDC), in Active Directory this role is held by the Domain Controllers.
• Client, these are the users.
• Server, these are the computers in your network.

With this basic division of roles, things work well in a domain, until a service account like that of the SQL Server database engine is a domain user. If so, that particular domain user should be made a server too. To mark an account as a service, a Service Principal Name must be registered. With NetworkService, this was done by the server (because servers can write SPN's on their own behalf), for any other account, by default only Domain Administrators can write an SPN. Use the SetSPN tool to mark an account as Service Principal Name, SetSPN can be downloaded, is part of the Windows Server 2003 Support Tools and is installed on Windows Server 2008 with the AD DS role.

To register an SPN for a service account, use the following syntax:

SetSPN -a MsSqlSvc/server1.domain.local:1433 domain\user

SetSPN -a MsSqlSvc/server1:1433 domain\user

Substitute the blue parts in the syntax for the server names (FQDN and NetBIOS), the port where SQL Server listens (this could be different from 1433, especially when registering an SPN for additional instances) and the domain user account used as SQL Server service account.

When running more SQL Server instances under the same domain account, it may be useful to check the approach listed in Step 3 of How to use Kerberos authentication in SQL Server, so the AD-people have to be called upon only once for the service account, not for every instance installation.

Reference Environment

As reference environment (may be a layout for a test-network), the following machines, users and services are used in a single domain configuration;

• Windows Server 2003 or 2008: DC1 (domain controller), SERVER1 (member), SERVER2 (member)
• Windows XP, Vista or 7: CLIENT (member)
• SQL Server 2005 or 2008: SERVER1 (default instance) tcp 1433 static, SERVER1\SQLINST1 tcp dynamic (named instance), SERVER2\SQLINST1 (named instance) tcp 1433 static

To make it work, user Domain\Bobbie has to be trusted for delegation in all four scenario's. Depending on the service accounts for SERVER1\SQLINST1 and SERVER2\SQLINST2, use the following settings

Things to consider

When working with Kerberos, tickets are granted for a longer period of time (typically 10 hours), so configuration changes usually take a longer time to apply. Use the KERBTRAY and/or KLIST resource kit utilities to view and purge tickets, in order to speed things up (avoids rebooting systems or logging off/on users). Alternatively, consider shortening the ticket lifetimes to the minimum; 10 minutes for services and 1 hour for users.

When configuring the linked servers, it was best to do so at the server where the linked server is configured from. Depending on your role you may be a domain admin, so should domain admins be delegated in order to make setting up linked servers easier? Rather not! Imagine someone installed malicious program or script on a system that is trusted for delegation and somehow a domain administrator executes this program or script. It could hurt any system or user in your network.

I knew this was coming and still I was taken by surprise; my son Bram was born last week (July 27th). An event like that turns your life upside down, forever. I'm not in the habit of putting personal stuff on the net, but this is worth the exception. So here they are, a proud father and his one day old son.

Sometimes SQL Server gets it wrong, UNIQUE constraints and NULL is one of the parts where it is wrong. People call it a bug and request for it to be fixed, Microsoft calls requests to fix it common and offers workarounds and overall current SQL Server behavior does not adhere to the standard (ISO-9075).

To see the bug reports (and vote); 126533, 299229, 311223 (closed), 387273. Basically they are each others duplicates and from those, the request 299229 has most attention and the best discussion.

Available workarounds;

• Filtered index (starting with SQL Server 2008)
• Indexed view (starting with SQL Server 2000)
• Triggers (not recommended as the check is after the modification is completed and if invalid has to perform a rollback)

... and don't say Microsoft didn't get it right because it's Microsoft... Access implements UNIQUE constraints correctly.

I've been holding this one off for quite some time, but finally bit the bullet... migrate the server.Coming from Windows Small Business Server 2003 (32-bit) going to Windows Server 2008 with Exchange 2007 (both x64), including transferring the web site. There were a few nasty pieces in the process and I didn't have/take the time/resources to test it all though before jumping in... so I got a disruption of about one hour for the blog and inbound smtp.

First issue I encountered was the installation of the server, which I ordered from Dell without operating system. I am a Microsoft Action Pack subscriber, so I slammed the 64-bit Windows Server from the Action Pack on the system, entered the Product Key and then went on to phone based activation. After putting in the last confirmation code and pushing the next button... my activation wasn't accepted!!! ... to cut a long story short, after dozens of voice menu selections and 4 persons, I found the guy who could help me out: Philip. Though it still toke Philip and me about 20 minutes to find the actual problem... I wasn't an Action Pack subscriber anymore!!! Technical story here, but it happened when Action Pack was integrated more tightly with the Partner Program. Somehow I didn't complete the Action Pack renewal transaction completely in February, but the partner site didn't show that (it only showed I'm still good for the Partner Program).

Couple of days later, I could go ahead with the migration process. It turned out to be easier than I expected after I found a very useful document about Upgrading Small Business Server 2003 to Exchange 2007. Some pointers on using the document though;

• Transferring the the other FSMO's (Schema Master is described); Determining FSMO Role Holders
• With Windows Server 2008 you NEED the Exchange 2007 SP1 DVD, Exchange 2007 without SP1 will not pass installation checks.
• Section Migrate mail send does not cover migrate mail receive. This caused me a bit of downtime... The issue was
  SMTP error from remote mail server after MAIL FROM:<abcd@efg.hi> SIZE=2703:
  host abcdef.ghi [10.10.10.10]: 530 5.7.1 Client was not authenticated
  And can be solved in the following manner in Exchange Management Console:
  • Go to the Server Configuration, Hub Transport. Select your server and add Anonymous to the Default SERVERNAME Receive Connector. While you're there, you can increase receiving message size here too.

And there were parts I did different;

• I continue to rely on Vamsoft's ORF for the anti-spam. Version 4.3 is fully compatible with Windows Server 2008 and Exchange 2007 and an ease to configure.
• Since I will continue the same DNS name for OWA and Windows mobile Active Sync, I transferred the Self-signed certificate from the SBS2003 (find it in your IIS, web site, directory security tab) to the new server.

My other bit of downtime was the blog... simply copy from the old server to the new one didn't cut it. So I downloaded the latest release of dasBlog from Codeplex, copied it into the virtual director of new server. Next I had to;

• Copy my own theme
• Compare (and adjust) the \SiteConfig files
• Transfer \Content files and \Logs files
• Make sure the account running the Application Pool has read on all dasBlog folders and change on \SiteConfig, \Content and \Logs.
• The Application Pool is running in Classic Managed Pipeline Mode

Guess I can say I didn't plan for the blog to be upgraded... that just had to happen.

Microsoft launched a new preview site for SQL Server 2008 R2. You can sign up for notifications about the releases of CTP's.

From an administrative viewpoint, focus seems at tools to manage, utilize and consolidate instances and machines.

Towards usability (mainly business intelligence), enhancements to integrate with Office 2010 (mainly Excel and SharePoint) are featured. Also further integration with Virtual Earth is shown.

Finally there is mention about something called Master Data Services and the Stewardship Portal. In concept a very powerful way to create and maintain a data warehouse bus.