Tag Archives: SQLPASS

Third day at #SQLPass, the real sessions begin #sqlblog

Finally the conference started for real. So many people all over the place. Crowded in all sessions and even the dining hall was almost full for lunch.

Today I attended four session. First was the Key-note with Bill Graziano from Pass and Quentin Clark from Microsoft. SQL 2014 CTP 2 was announced and and brief walk through of all new features and  functionallity was very interesting. In memory databases was demonstrated and this looks cool. Huge improvments will be availabale by just altering a table or procedure to take advantge of this functionality. No code rewriting is necessary.

Power BI Q&A was also a new feature that looks very promising and the audience seem to like it very much.

SQL backups to Azure storage also seemed like a nice feature, especially since it’s available also for 2005 and 2008 via a downloadable tool. There was also a new feature that would automatically trigger transaction log backups based on the amount of changes data in the tables.

The keynote is available on http://www.sqlpass.org/summit/2013/Live.aspx

After that there was a great session with Erin Stellato about “Taking the leap from Profiler to Extended Events”.  Very good presentation and it really made me understand that there are so much more you can do with Extended Events then with Profiler. Need to start using that when I get back home again.

Then there was a session with Glenn Alan Berry called the “Professor of DMV”, where he went through his pack of DMV scripts that can be used to troubleshoot SQL server performance problems. I have used these DMV script before but I have never really fully been able too utilize them to it’s full potential. Now I know how to use them better and will dig into them again. They can be downloaded here:  http://sqlserverperformance.wordpress.com/tag/dmv-queries/

Then I attended a seesion about automating your SQL server deployment with Powershell. The tool is called SPADE and is available on http://www.codeplex.com. I have started to automate SQL server installations myself but my still limited knowledge of Powershell prevented my from going all the way. This tool will save tons of time spent on installing and configuring SQL server.

Lats session was about Troubleshooting Clusters and Alan Hirt gave a nice walk through of some pitfalls in cluster install and configuration and how to troubleshot them.

Then we attended the Exhibition recpetion for some food and chat and ended up on a pub for a couple of beer.



Make your SQL Server Apps go faster, day 2 pre-con at #SQLPass

Second day pre-conference at SQL Pass Summit for me was a session with Brent Ozar, Kendra Little and Jes Schultz Borland called “Make your SQL Server Apps go faster”

Here is some of the things that we learned today.

Key focus in the presentation was to see what can be done to make applications run faster from a DBA point of view. What can I as a DBA do without involving a developer or changing any of the application code.

First we need to know what is our worst performing queries, there are a number of ways to find that out. Activity Monster Monitor is NOT the best tools for this, beacuse it’s not telling the truth all the time.

sp_whoisactive is a free stored procedure that can be downloaded and installed. It can be downloaded from here http://sqlblog.com/files/folders/beta/entry42453.aspx. Brent Ozar also has a quick guide here  http://www.brentozar.com/archive/2010/09/sql-server-dba-scripts-how-to-find-slow-sql-server-queries/.

Kendra Little aslo gave a quick guiding to Extended Events and how that can be used to give us information of our top queries, more about that in my next post after I have attended the “Making the Leap from Profiler to Extended Events” tomorrow.

We are also interested in what kind of waits we have in our server, this together with the information we get about the heaviest queries will help us determine what choices we have to tune.

SQL server have some DMV’s that can be queried to find missing indexes, there is also the Database Tuning Advisor that is commonly used by developers to create indexes on tables. Netiher of these are very good tools, just implementing recommendation from theses tools might give you worse performance in the long run. Recommendations should really been seen as just that. It is vital to know how your tables are accessed and queried to be able to build the correct indexes.

Correct indexing can not only solve direct IO problems but can in some cases also help to solve  fixed implicit conversions.

Sp_blitzindex is a great tool to figure out your index needs. It available here http://brentozar.com/go/blitzindex. It will also help you find your duplicates.

It’s also important to know that some simple queries are executed as trivial and will not generate a index recommendation. If they are run frequently it might be worth while to look at it anyway. Find them with sp_whoisactive.

Remember that in SQL 2012 index recommendations are flushed when a index rebuild takes place, this was not the case in earlier versions. Index defragmentation does not flush the recommendations, this can make it tricky to maintain indexes in SQL 2012.

Index maintenance should be carried out every 4-6 weeks, this is because usage pattern and data volumes changes a lot over time.

CX_PACKET is a wait stat that is very common in SQL Server and often it is thought of as a problem. A common reason why this can become a problem is the “parallel threshold” value in the SQL Server configuration. Default value for that has been 5 since a very long time. In todays high performance servers it is way to low. A starting value of 50 is now days the recommendation. MAXDOP is also taking part here and that should not be left at default value of 0. Less or equal to number of cores on server is a good starting point. Usually no benefit over 16.

Always try to avoid table variables unless your 100% sure they will ever return more then one row. http://brentozar.com/go/serialudf describes this more in detail.

These where just some takeaways from todays pre-con. It will be interesting to get home and practice this on real data…..



SQL Pass Summit 2013 first impressions

Second day of pre-confrenece at SQL Pass Summit is about to begin and I have not yet summarize yesterday.

First day turned out to be very interesting. My first pre-conference session, “From SQL Server Reporting Rookie to Rock star”, maybe didn’t make a Rock star in building Reporting services reports but it was a very good walk through of all the possibilities within in the product.

Stacia Misner did a very good job describing all the functionality. But I got a feeling that she felt that a lot of focus have been redirected towards PowerView and integration with Sharepoint and not in developing Reporting services. Nothing much really was new in 2012 and it seems there is limitied news in 2014 as well.

Becoming a Rock star in Reporting Services will for sure require a lot of reporting writing.

In the evening there was a Networking dinner at the Buffalo Wild wings where @Goranpeterson and I had a long talk with @DavidAMacLean1. He owns a castle in Scotland is related to Sean Connery is some distant way and a big fan of action movies.

Still not a lot of people here, most badges are still waiting in registration, if think it’s going be crowded here tomorrow.



Planning my #SQLPass session schedule….

Started planning my session schedule for SQLPASS while I was waiting inline for immigrations at O’Hare airport. It was a long line so I had time to go through most of the details in the #Guidebook and I picked everything I liked.

No I have a schedule with 3, 4 or even more sessions I want to go to for any given time. How should I choose ?  As a #sqlfirsttimer I’m not sure what the best approach is ? Should I go for the   speaker that I have read and listened to on webinars, youtube, the once that I’m following in Twitter and who seems to really know their stuff ? Or should I go for the once that I never heard off and hope that they have something completely new to teach me ?

There are so many interesting areas I would like to dig inn to, should I choose one area all the way or should I mix them as much as possible to get new input ?

BigData, SANLess Clustering, Extended Events, Power BI and Power Map, database corruption, AlwaysOn, Powershell, PDW, 2014 InMemory OLTP, Clustered updatable columnstore indexes, Indexes, Central Managment Server, SQLCAT session, SQL Clinic, Query optimising, etc, etc…

Not to mention all the other activities, First-Timers orientation, Welcome reception, Exhibitor reception, #SQLRun,  Sponsor breakfasts, Keynote, Community appreciation Party at Nascar hall of Fame, Birds of a Feather Luncheon, SQL Server Clinic an SQLCAT, SQL Kilt day and #SQLKaraoke.

I starting to believe I will not make to all of these events. But I will sure try. I will probably end up bying the DVD…