Wer noch keine Idee hat, worum es sich bei dem SQLSaturday handelt, der guckt sich kurz meinen vorangegangenen Blog Post an und natürlich die offizielle Seite www.sqlsaturday.com und dann SQLSaturday #230 Rheinland.

Der so genannte “Call for speakers” findet ja auf der Eventseite statt. Wer also gerne eine Präsentation zum Thema SQL Server (DBA, BI, DEV) halten möchte, der kann einen Session Vorschlag einreichen.

Ich mache hingehen den Ruf nach Fotografen. Warum? Nun, ich bin Anfang des Monats von einem Organisator (um genau zu sein: von Tillmann (b|t)) des ersten und auch des kommenden SQLSaturdays in Deutschland gefragt worden, ob ich nicht dieses Event mit der Kamera begleiten möchte.

Da ich ganz gerne mal zur Kamera greife, werde ich natürlich das Event, so gut es geht, in Fotos festhalten.

Da es an dem Tag mehrere Session Tracks geben wird und ich nicht in jeder Session zugegen sein kann suche ich ein wenig Unterstützung.

Konkret ist Folgendes geplant:

· Fotos aller Sessions

· Fotos während der Pausen des Events

· Fotos der Sponsoren

· Fotos von den Preisverleihungen (Die Sponsoren, welche mit Ständen vertreten sind, bringen gerne mal das ein oder andere Gimmick mit)

Optimal wäre es, wenn sich 2-3 Leute (gerne auch mehr) finden, die ebenfalls Spaß haben, den SQLSaturday mit der Kamera zu begleiten.

Zwei wichtige Punkte muss ich aber noch erwähnen:

· natürlich sollt Ihr auch von der Konferenz an sich etwas mitbekommen. Aber je nachdem, wie viele Freiwillige sich finden, kann es sein, dass Ihr nicht jede Wunsch-Session mitnehmen könnt. So oder so muss man hier aber noch den Session Schedule abwarten. Vielleicht passt es ja auch auf Anhieb.

· die Mitarbeit erfolgt unentgeltlich. Die SQLPass lebt ja primär von der freiwilligen Mitarbeit, daher kann Ich Euch kein Honorar anbieten. Die Leute, die mitmachen, würde ich dann aber schon auf ein Bierchen oder Ähnliches einladen

Zum Thema Equipment: Es ist keine High End Ausrüstung notwendig. Eine Kamera mit einem Standard Zoom und Blitz (oder ein lichtstarkes Objektiv) sollte passen. Wer im Besitz eines leichten Statives ist kann dieses auch mal mit einpacken. Freihandaufnahmen in etwas dunkleren Konferenzräumen sind nicht immer so einfach.

Wer also den SQLSaturday besuchen will und auch noch Spaß daran hat diesen in Fotos festzuhalten, der melde sich doch einfach kurz bei mir. Sei es hier als Kommentar oder per Email, Twitter, Xing, FB.

For me it was also the first time that I not only attended such an event, but also volunteered as well.

So, for me the event already started on Friday where I helped with the preparations. The main task was to pack the SQLSaturday bags, which was really time consuming. On the other hand I still had time to take a few pictures of the preparations. To get an idea, please take a look here.

The Saturday itself was, in the beginning, a bit stressful for me. Why? Well, I also helped at the registration (e.g. handing out the swag bags, badges, t-shirts)  that morning and the people came in waves. But hey, when the first sessions started at 9 a.m. everything was fine and I could also attend (with a little delay of 15 minutes ) the first session that day “Execution plans in SQL Server” by Michael Rother .

Then it was Andreas Wolter’s turn, talking about Tracing with Extended Events. Really an interesting session and a requiem for SQL Server Profiler (“Hasta la vista, Profiler”). Andreas combination of his knowledge and humour made it fun session.

Without going in too much detail, just a little sum-up of the other sessions I have visited: SSIS Change Data Capture hosted by SQL Server MVP Allan Mitchell (twitter). This session was well-attended and and a good mixture of slides and demoes.

After the lunch break (pizza

time!), I just made a quick visit to Oliver Engels and Tillmann Eitelbergs session, when they were talking about Data Quality Services, followed by Hugo Kornelis’ (blog|twitter) “Everything you always wanted to know about MERGE” session which also was one of my highlights (UPSERTS,DUPSERTS and “we get the maximum of two…. things”).  After a little photo session break where Tobiasz Koprowski and I took some pictures of all speakers it was time for the last session of that day. I went to the “Introduction to (MERGE) Replication” session by Johannes Curio.

But the SQLSaturday did not end with the last session. There were some raffles since the sponsors had a lot of prices to give and I have to say “Thanks” to Quest Software Germany for the Amazon voucher I got .

To wrap it up: The first SQL Saturday in Germany was a great experience for me. I have met many people I only knew through twitter or their community work. Just to mention a few: Torsten Schüssler, Niko Neugebauer, Hugo Kornelis.

Thanks to all the other Volunteers, it was a lot of fun working together and hopefully not the last time.

Last, but not least, thank you to all the speakers who spent their time to present at the SQLSaturday. What an awesome crowd:

A very special thanks to the most important persons of this SQLSaturday: Kostja Klein, Oliver Engels and Tillmann Eitelberg, who invested a lot of time to make this event possible.

P.S. You find all of my pictures from this event here.

Jen asked a lot of interesting questions and I have to confess, I couldn’t really answer all of them without hesitation. One of her questions/tasks was to name three DMVs you use on a regular basis.  Luckily I wasn’t the only one who said: “I can’t”. It was Brent Ozar (blog | twitter), the well-known community-contributor, MVP and MCM.

So, his reply regarding the DMVs was:

Name three DMVs you use on a regular basis. – I can’t. I’ve had the same scripts in use for years, and even though I use it at least three times a day, I can never remember if it’s sys.dm_db_index_usage_from_behind or whatever. That’s what script files are for. If you’re typing the DMV name by hand multiple times per day, you’re doing it wrong.

This answer made me feel a bit better and I got the idea to write this post.

To answer now Jen’s question: I cannot name the three DMVs I use regularly, but I can name the three scripts/procedures I’m using. These are:

• Glenn Berry’s Diagnostic Information queries (you’ll find them on his blog)
• Brent Ozar’s Blitz Script
• Adam Machanic’s Who is Active

To have Glenn’s script quick at hand or to call the two procedures immediately, I use the SQL Snippets functionality from the SSMS Tools Pack by Mladen Prajdić (blog|twitter). For me, this is one of the most useful features, since I can call complete code blocks just by typing in a shortcut and hit enter.

If you have the SSMS Tools Pack installed, just go to the SQL Snippets option

Here you can define your snippets and put your sql code, custom scripts etc. pp. in it.

If I now want to open the Diagnostic queries in my query window, I just type glenn28_new and I get the whole script:

I’m calling the Blitz script and Who is Active in a similar way.  Wia for example just opens
EXEC sp_whoisactive,

but with wia2 I get

EXEC sp_whoisactive @show_sleeping_spids = 2

This is just a quick example how I call my scripts I use on a regular basis.

today I’ve spent little time installing the new SQL Server 2012 Performance Dashboard (you can find it here).  Then I picked a test server (still SQL Server 2008 R2, build .2500) and ran the setup.sql to provide all the necessary procedures for the reports.

When I tried to open the “Recent_CPU” Report by clicking on the blue bar

I got the error “A data source instance has not been supplied for the data source ‘CHECK_DEPENDENCIES’. After a refresh of the report a 2nd error showed up: Error converting data type nvarchar(max) to datetime. My first thought was: “Hey, post a quick question at ask.sqlservercentral.com and wait for answers.”  Then I got ambitious and started SQL Profiler. Since it was only a quick trace I ran the GUI. Normally I would have created a server-side trace. You still don’t know what a server-side trace is? Take a look at Stairway to Server-side Tracing.

I catched a user error saying “Error converting data type nvarchar(max) to datetime.” right after this statement:
EXEC sp_executesql @stmt=N'exec msdb.MS_PerfDashboard.usp_QueryStatsRecentActivity @WithActivitySince',@params=N'@WithActivitySince NVarChar(max)',@WithActivitySince=N'27.07.2012 19:52:17'


That was an interesting information. So I opened the recent_cpu.rdl in BIDS2008 and checked the Parameters and guess what? The parameter ‘WithActivitySince’ was declared as string. Since I already made my experience messing up reports with BIDS 2008 when the origin report was created with the 2005 XML schema I just opened the rdl file with notepad++ and changed the defined data type  from “String”

<ReportParameter Name="WithActivitySince">

 <DataType>String</DataType>

  <AllowBlank>true</AllowBlank>

  <Prompt>WithActivitySince</Prompt>

</ReportParameter>

to “DateTime”

<ReportParameter Name="WithActivitySince">

 <DataType>DateTime</DataType>

  <AllowBlank>true</AllowBlank>

  <Prompt>WithActivitySince</Prompt>

</ReportParameter>

After this little change I was able to run the recent CPU report without any problems and double-checked that the report still worked on a SQL 2012 as well.

Regards and have a great weekend

this is jut a quick thought (or a reminder) regarding “orphaned users”: There are several scripts out there on how to find an orphaned user account inside a database (I just stumbled across a script a few hours ago, that’s why I’m writing down these words).

Most times DBAs have to deal with an orphaned user when they restore a database to a different server. What then happens is that the SID of the user account won’t match any longer the one from the “old” server. Therefore you have to “re-map” the database user to a valid SQL Server login.  For further information take a look at BOL.

But why am I writing stuff you all already know? Well, there are a few scripts out there which do not only find orphaned users for you but also gives you the chance to delete them.  And now my reminder: You have to double check if the database user really isn’t used.

Do you’ve ever heard of a so called “loginless user”?  This is a user which only exists inside a database (e.g. there is no corresponding server account). Maybe this user is used as a replacement for an application role? That means a user connects through a application with his credentials to SQL Server, the app then initiates an “Execute as” and changes the context to our loginless user. If that’s the case then it wouldn’t be the best idea to drop that database user….

So be aware:  loginless user != orphaned user

For native SQL backups we have no dedicated disk space on our database servers. Instead we´re running native backups over the network and place them on a UNC path. I know that it is probably not best practice but it fits for our environment (or to quote André Kamman (blog | twitter): “It works for me!”) very well.

It is Murphy´s law that sometimes things can go wrong. So it happened on a few machines that the backup was created but finished with an error:

The operating system returned the error ’64(The specified network name is no longer available.)’ while attempting ‘FlushFileBuffers’ on ‘\\OurBackupServer\ourshare$\OurServer\OurDatabase\FULL\ OurServer_OurDatabase_FULL_20111021_041523.bak’

So we did a little bit of research and even talked to Ola Hallengren when we visited SQLBits 9. His assumption was some kind of timeout.

Since this error only occurred on some test machines we didn´t spent too much time investigating / solving the issue (we could live with it) until it happened also in a productive environment.

Then we came across a msdn forum entry (take a look here) and tried the mentioned solution to create a DWORD with name ‘Sesstimeout’ and value 360 in  HKLM\SYSTEM\CurrentControlSet\Services\lanmanworkstation\parameters.

And that was the solution. Since we´ve modified the registry on the  involved SQL Servers the backups running fine without any issues.

this is just a quick blog post as a short reminder regarding side by side installations. I just want to share my little experience I made yesterday. Maybe it is obvious for most of you, for me it was a little surprise.

Most of the database servers I support are SQL Server 2008 installations. Some of these also have snapshot replication enabled.

To prepare the launch of a new application / migration of an existing application we made a side by side installation. Beside an existing SQL Server 2008 there´s now a SQL Server 2008 R2 as well. And guess what? After the installation of  SQL 2K8 R2, replication jobs were failing on the SQL 2K8 instance. Also, in the job history of the corresponding SQL Agent Jobs there was an entry about a missing file. After a little research I figured out that the distrib.exe was no longer available.

So what happened? The R2 instances where installed without the SQL Server Replication components since these are no longer needed for the new application. So the installer  (for whatever reason) removed the existing replication components which where used by the SQL 2K8 instances.

The good thing is that I could easily reinstall the Replication components and now the replication jobs are running fine again.

Special thanks go to:

Ted Krueger (@onpnt) and Jose Chinchilla (@sqljoe) for the good ideas arguing against running SSAS on a productive OTLP system.

Aaron Bertrand (@AaronBertrand),  Grant Fritchey (@GFritchey), Jonathan Kehayias (@sqlsarg) for the good hints regarding my Performance Counter question.

A very special thank-you goes to David M Maxwell (@dmmaxwell). David made a whole blog post to answer my question. I was really impressed.

Also, thanks to all the blog authors I read regularly (and also follow on twitter):

Brent Ozar (@BrentO) – his posts are always very good to read and we all Trololo with BrentO (SCNR)

Paul Randal (@PaulRandal) – MUST READ if you work with SQL Server (the whole sqlskill.com site is worth visiting)

David Levy (@Dave_Levy) – A lot of useful stuff for the day to day business. Also David told me how to use Twitter properly regarding the information about new blog posts.

John Sansom (@JohnSansom) – I like his “Something for the Weekend” posts very much. Catched a lot of interesting articles in the past which I would have truly missed if John did not do this overview.

Hopefully I will have one or two great ideas by myself which I can then share with the community as well, just to give a little bit back.

Regards,

Dirk

I have to apologize, because I made a little mistake in my previous post.

It is correct that the MBCA has a problem running properly when group policies are defined, BUT I was wrong with the remote scanning.

Let me explain my little “mistake”:

If you´re running the MBCA there are two ways scanning a machine.

1. choosing ‘Connect to another Computer’ as shown here

Then you´ll probably get the error as described in the previous post (the execution policy stuff)

So, the 2nd way to go is:

Choose the SQL Server 2008 BPA from Dropdown menu and click ‘Start Scan’

In the next menu, you can define parameters for scanning as well as another computer!

So, sorry that I was wrong saying you cannot scan remote computers easily.

I will try to play around a little bit with the BPA. If I find more interesting stuff or issues, I will let you know.

Here is my little follow-up to the blog post yesterday.

I have moved the machine with the installation of  MBCA and BPA into an OU where the inheritance of the domain policies is blocked. I have checked the PS execution policy with Get-Executionpolicy –list . Every scope was set to undefined except UserPolicy…. I forgot to place my test account into the test OU without the policy inheritance.

Now, that I had a clean system, I could run the MBCA without any problems and therefore the SQL Server 2k8 R2 BPA as well. But once I have enabled a policy setting, either for MachinePolicy or UserPolicy, MBCA failed again. Keep in mind that in the environment I want to use the MBCA/BPA, the policy regarding the PS ExecutionPoliciy is set to unrestricted.

Another problem was trying to scan a remote machine with the MBCA. If you run the MBCA locally it will call the MBCA on the target machine which then loads the BPA there. In my case it has failed since the target server also gets the ExecutionPolicy setting via a group policy.

What I will do next is to check if this behaviour still occurs if the ExecutionPolicy is set to RemoteSigned by a group policy.

And a last little hint: If you run the BPA through Powershell and you want to scan a SQL server instance, either local or remote, use an account which is a direct member of the sysadmin group. I used an account which is member of a special domain group and this group then is member of sysadmin-group on the sql servers. In this scenario I got only one error telling me that the account I´m using is no member of sysadmin group….