I’m Speaking at PASS Summit!

On May 31st, I was shocked to find out one of my submitted sessions was selected for PASS Summit 2018. My response?

I didn’t see this coming. At all. I was stunned – walking around at the office, saying “Oh my gosh…I got picked” while co-workers looked at me like I had lost my mind. Literally.

I thought I would be going to PASS Summit this year, but not as a speaker. I thought I would be an attendee, learning a lot and having an amazing time. I thought it was far more likely that we see 10 feet of snow this winter in Texas than I would be selected for this year’s PASS Summit.

I am sincerely humbled and grateful to be selected. The competition was stiff. I should know – just like last year, I pulled all the data while it was still available. Once the selections were announced I was able to isolate the sessions from those submitted, and also isolate the ones that were not selected. I was able to determine one particular number – how many sessions were selected out of total submissions.

90. 90 sessions were selected out of approximately 600 submitted. I might be off by a few, but I also checked this against ones that are there but I didn’t have in the data pull from immediately following the closing of submissions. For the most part, the ones that are there that didn’t match with a submission are pre-cons or sessions from what I am assuming were invited speakers.

I’m not saying any of this is bad. What I am saying is that the competition among PASS Summit submissions is growing and becoming more fierce. I see this as a positive for the community – we are developing more and more excellent speakers with awesome content to share.

I queried my data set further, looking for the names of the people selected and those not selected. I found the latter to contain many speakers I know personally, who would make amazing PASS Summit speakers, and may have already done so in the past.

I pulled the list of the speakers that had sessions that were part of the “Best Of” from PASS Summit 2017. Out of 36 speakers from that list, 13 of those are currently part of 2018 PASS Summit lineup. While it could be said that all of these people should be on this year’s roster because they were part of the “Best Of” from last year, I think the only thing that can really be taken away from this number is the fact that no one is guaranteed a speaking spot at PASS Summit, no matter who you are.

Are there invited speakers? Yes, but like the regular submissions selection, that list likely has some subjectivity to it as well. Maybe an invited speaker gets red carded in the previous year following their session (heaven forbid, but you never know) – do you think they would be invited back? I would hope not, regardless of the content they would provide.

As speakers in the SQL community, we are largely a very goal oriented, driven bunch of people. We might even be a tad competitive. It’s not lost on me how fortunate I am to have been selected this year. Looking back to previous years when my submissions didn’t make the cut, I know now that while this was something that I wanted, I wasn’t ready. I might have been grumpy about it, but that didn’t make me any more ready for this monumental task. For all the seasoned PASS Summit speaker veterans that were not selected, you of all people know how competitive this is. You also know that your time will come again.

Selected or not, I hope to see everyone at PASS Summit this year. If you haven’t registered yet, and need a discount code, I got one of those for ya!
LSDIS97QR

The price for PASS Summit goes up after June 29th so if you are planning on attending, get registered now!

Note: I am on the board of directors for the North Texas SQL Server User Group and this is our code. Registering with this code benefits our group. There are other discount codes out there but I just gave you this one. Like, right now. See? There it is. So go register and use it!

Being a Speaker

Back in 2013 I attended my first PASS Summit. I was a first timer and Bill Fellows was my “big brother.” I had no clue what I was getting myself into. This is my story on how I became a speaker.

At my first summit, thanks to Bill, I was meeting people just about from the moment I got there. Bill knew lots of people. He’s tall and bald and hard to miss. And wears shorts. Everywhere. Regardless of temperature.

Where was I…oh yeah…speaking. While at summit that year I met and talked to many people. A few of them asked me “Have you ever thought about speaking?” I politely replied that I hadn’t, but I really wanted to ask them if they had forgotten their medication that day, or if they needed to be examined by a professional. Me – a speaker. LOLOLOLOLOLOL! Bless their hearts!

But then I thought about it. Why not speak? What would I talk about? What was something cool that I had done that people would want to know about? Could I do this? I showed people how to properly gift wrap a present in speech class (oh yes I did…and there is a right way…and you’re probably doing it wrong) – could I teach them something that would actually have some career benefit?

Yes. Yes I could.

I started to think about what I could talk about. I submitted a session to a SQL Saturday event later in 2013…and I was not selected. I didn’t know it at the time but I had picked a topic that was rather niche and would not have a real following. I found this out when I went to the event (yes, I still went…I was going to this one regardless) and talked to the organizers. This also allowed them to know who I was, so when I submitted again, they would be able to put a face with the name.

The following year I submitted a panel to a different SQL Saturday…and they picked it! This allowed me to get my feet wet while having some other more seasoned speakers there to help drive the conversation. In the weeks leading up to the event, we met and ironed out the details of what we would talk about. During those meetings I learned a lot from them, and they helped make that first session for me a success.

As a speaker, there are good days…and not good days.

A few months later I did my first solo session – and it did not go well. This was a tough pill to swallow. What happened? Simply put, my session didn’t reflect my abstract. I was disappointed in myself. It takes a few days for me to get over things like this, but since this was my first solo attempt, it really made me question whether or not I wanted to do this. Could I be good at this? I went home and re-wrote the session. Then looked at it and re-wrote it again. The next time I gave this session it went much better, and my feedback reflected it.

I love things that challenge me. Speaking does this, and does it in ways I would have never imagined. It takes me out of my comfort zone. No matter where I am for the foreseeable future, I will be speaking.

Speaking has taken my love of learning new things to the next level. Not only am I learning for the sake of a problem or making a business case for something, but now I am learning with a goal or being able to facilitate the learning of others. When I refine my skills, my session material gets better too.

My passions have found an outlet with speaking. It’s thrilling and sometimes frustrating. When you are speaking you never know what’s going to happen. Sometimes you have a lot of questions from the audience. Other times you have none. Sometimes you don’t know the answer. As much as you practice, it doesn’t change the fact that as many times as you give a certain session, no two of those are alike because no two audiences are alike. The one thing that is consistent is that the more times I am able to speak, the more people I am able to reach and help improve their SQL Server skill sets or adopt a new skill set with PowerShell. I don’t know where speaking is going to take me, but for now, I’m down for the ride to find out.


If you are interested in speaking, and would like to find out more, here are a few blog posts I found on this subject.

Paul Randal (via Grant Fritchey) – I think this one lays it out pretty good.

Cathrine Wilhelmsen – A really good post on when things don’t go as you had hoped.

Thomas La Rock – Once upon a time, he was a mere SQL fanboi.

Kevin Hill – On starting to speak and then after his first time speaking at SQL Saturday.

These are what I could find. If you have a blog post on speaking leave a comment and I can add you. 🙂

T-SQL Tuesday #100 – Predicting the Future!

For this T-SQL Tuesday we are asked to look into our crystal SQL Server ball and predict what will be happening at the time of T-SQL Tuesday #200. I went to the garage, dug that thing out, cleaned it up, and boy it had a lot to say!

Assuming we’re all friends here, and there is some fun to be had with this…

T-SQL and JSON had a baby. All queries in SSMS resemble a hybrid of the two languages.

SSDT has been replaced by VSDT. Nothing has really changed but the acronym. As always, you can still expect some things to break when you do updates.

Microsoft bought NHibernate. You still have all the same issues as before but now you post them to Microsoft Party (it replaced Collaborate…after that replaced Connect) and actually watch them not get fixed. And you can’t post work-arounds in MS Party (so it’s not much of a party).

MS NHibernate still generates SQL queries that are long and redundant, but it’s not handling the TSQL-JSON baby very well. So there’s that.

Microsoft acquired ActiveBatch and it is now called SQL Server Batch and has replaced SQL Server Agent for scheduling jobs in SQL Server 2026. Consequently, companies have been reluctant to upgrade from SQL Server 2023 (especially the ones that have used ActiveBatch).

For the companies that are upgrading, they have found that calling PowerShell scripts from scheduled tasks to be a good way to bypass using SQL Server Batch. Increase the in the demand for DBA’s with extensive PowerShell experience sky rockets!

The rumors back in 2018 proved to be unfounded – DBA’s are still in high demand. All the talk of SQL Server tuning itself turned out to be DTA 2.0.

Microsoft brought back the MCM. And then killed it again the next year.

Azure has been replaced by Rainbow. Data is no longer in the “cloud” – it is in “rainbows.” Pricing is based on the colors of the rainbow and the color names are garnet, citron, lemon, lime, azure, and violet.

PASS still exists. Due to some bylaw changes, elections have not been held since 2019. Grant Fritchey [B|T] is still president and attends meetings remotely from his nursing home.

Just kidding – Grant’s not in a nursing home. That’s just where he says he is. There was some backlash when PASS did away with SQL Saturday events. Grant’s really in witness protection and goes by the name Thomas LaRock [B|T].


Thanks to Adam Machanic [B|T] for hosting the T-SQL Tuesday this month, and for coming up with this whole thing to inspire all of us to write more and continue to share knowledge. While there was absolutely no knowledge in this post, I do hope that I got a giggle from at least one person.

 

Quick and Dirty UPDATE STATS Job with T-SQL and PowerShell

When performance issues rear their pretty little heads I tend to look at the simplest things last. I am the person who is looking at the code first – because in many cases, if your code was better we wouldn’t be having this discussion. Whether it’s a series of nested views or one single user-defined function that doesn’t play well with the optimizer, there is usually something in the code itself that is contributing to the performance issue.

But what about the times when it isn’t the code or you can’t touch the code (I’m looking at you NHibernate…and all of your red-headed step-child offshoots and fellow ORMs). What about the cases when the poor performance situation only occurs periodically, or once a week on the same day, every week, and while you can see the slow-downs in your queries, and all of your waits and plans, the source of the problem isn’t any particular smoking gun – it’s lots of guns, firing off at random times, and the casualties are in your interfaces that depend on the data from the queries.

The one thing I don’t immediately go to is UPDATE STATS – and I should be looking at this. It’s something easy to run and if you still have issues, you have ruled it out.

Over estimation on rows = bad times for the optimizer.

In this case, I ran UPDATE STATS on one table in particular and immediately saw a marked improvement in my test query.

Ahhhh….much better.

This morning I ran the same query and things were back to where they were with performance. This was somewhat disheartening, but the troubleshooting must continue.

But what about auto-update stats? Is this on? Yes, but auto update statistics is only going to update based on a percentage sample of the records in the table. If your table contains millions of records, you may not be updating your statistics sufficiently to see an improvement. I am updating statistics with a FULL SCAN of the table. For larger tables, this may take some time, but you may also see a difference in your execution plan versus only updating statistics based on a sample.

In an effort to rule out whether or not statistics are definitely a factor, I want to UPDATE STATS on all the tables in my query, and at a specific time – sometime the day before we expect our slowdown to occur. I also want to be able to easily see how long the process ran and the duration of the update for each table. I could write all this to a table, and maybe I will do this later, but viewing this from job history is what I want right now – it is easy to look at and easy for everyone else to find and see as well.

Creating a job with multiple steps can sometimes be a bit painful and tedious, if the job is going to have A LOT of steps. With PowerShell, a server name and a database, I can dynamically create this job.

The result from this is a job that will update statistics on your list of tables, one at a time. All you have to do from there is give it a schedule.

All those steps…courtesy of the magic that is POWERSHELL!

For this script, I was getting my list of tables based on a specific query that was problematic. If you are doing this for a stored procedure, you can query for the list of dependent tables instead, and instead of hard coding the tables in this code, pass in the name of your stored procedure and get the list that way.

I hope if you have a situation where some one-off stats updates are needed you find this script to be useful. As always, comments and feedback are appreciated. 🙂

 

Last known good CHECKDB…for ALL your databases!

This morning Wednesday I walk into the office and immediately hear that CHECKDB is the source of issues on one of the servers and is the reason behind some errors that have been happening. While I don’t think this is the case (it might look like it on the surface but there is something else that is happening that is the actual cause) I also wanted to find out what CHECKDB was running at the time the errors occurred.

I needed information on when CHECKDB ran for each database. When you look for what you can run to find when CHECKDB was last run you find this blog post and also this blog post on grabbing this info. While these were very informative, they were for one database at a time. I need this for all the databases so I can try to not only find out when each one ran, but also use these time stamps to figure out the duration.

I took the code from Jason’s post and made a few changes, running this for all the database with sp_MSForEachDB.

The results gave me some good information to go on – the database where the errors occurred was not the same database where CHECKDB was running during that time frame. There were also some FlushCache messages in the log – some suspect that this is also tied to the errors that are happening. Since all this ties back to NHibernate code, I suspect something else is going on and further digging is needed.

Situation developing. 🙂

Blocking – I Saw What YOU Did and I Know Who YOU Are!

When I was growing up I remember my mom talking about an old, scary, movie that she saw when she was young. In the movie some teenagers were making prank phone calls saying “I saw what you did and I know who you are.” One of the calls happened to be made to a guy that just killed his wife. Joan Crawford played a woman that was romantically inclined to said murderer. She eventually meets her demise when he stabs her because she knew about the first murder.

How every DBA feels when there is blocking

While blocking in SQL server might not be a felony offense (it isn’t…but it should be – WHO’S WITH ME?) as the DBA you not only want to know what is being blocked, but also who is doing the blocking and what in the H-E-Double-Hockey-Sticks they are doing.

At SQL Saturday Orlando I talked about this very thing and the query I defer to for the information.

IMA GONNA HUNT U DOWN!

It might look complicated but it is actually very simple – query sys.sysprocesses with a cross apply using the sql_handle to get the text of the query, and then an outer apply with the same query again but you are joining to the blocking spid so you can get the text for the query that is doing the blocking. Beyond that, you can filter on various columns and refine your output

WHY ARE YOU BLOCKING YOURSELF? WHY ARE YOU BLOCKING YOURSELF? WHY ARE YOU BLOCKING YOURSELF?

Of course, I know you can’t run this without the code (and I know that’s why you’re here…because I SAW WHAT YOU DID!)

If you are looking for a good way to troubleshoot blocking I hope this helps. If you have some folks running queries that are making you stabby, run this, find out what is going on, and then remove their access try to help them so they aren’t making you stabby any more. Then tell them they need to buy you a beverage because they are still alive.

UPDATED: PowerShell – WE HAZ YUR ACTIVE DIRECTORY RIGHT HERE!

I discovered something last week – I had not blogged about little things that I thought I had blogged about. What the heck does that mean? It means that I tried to reference my blog for something because I thought “I totally blogged about that”…and found out that was not the case.

Starting now, I am fixing this situation. There was something that popped up today that called for a PowerShell script and the Get-ADGroupMember cmdlet – get a list of users from a list of groups. Some users are in there more than once so this needs to be a distinct list, unless you are into manually cleaning up things like this, and then I will be sad for you. Because that is kinda sad.

I originally wrote a script with two arrays (one for the initial list and one for the de-duped list of users), but even though this is quick and dirty, that was a little too dirty. Enter the Group-Object cmdlet – it takes this list of names and groups them. No black magic this time. Just a cmdlet, that comes baked into PowerShell giving me what I need.

What? You wanted the code too? Oh, OK.

There you have it – quick, dirty and to the point. Enjoy. 🙂

UPDATE: Mathias Jessen tweeted a one liner for this….so no need for the one array! Woohoo!

I was trying to do this but was also just trying to get it done, and if in doubt, I slap things in arrays. Thanks Mathias!

 

It’s About Time for…#SQLSatOrlando!

Next week I am returning to Orlando not only to visit the mouse at his house, but also to speak at SQL Saturday Orlando! Last year was my first time there and I had a great time. This also happened after the event was pushed back to November 2016 because of Hurricane Matthew.

This year I will talking about deadlocking and blocking – something that is an issue for so many DBAs. Whether this issue is indexes or bad code…or you have been trying in use indexes to cover up bad code, blocks and deadlocks can happen in even the best environments under the right circumstances. If you are going to be at SQL Saturday Orlando come to my session where we will talk about detecting and preventing these arch enemies of the DBA!

AzureRM, PowerShell and Deleting Your VM the Easy Way

Before I worked on server migrations to Azure, I had only worked with Azure for a SQL Saturday session, and I was only using Azure as a host to the virtual machines (VMs) that I used to demonstrate PowerShell scripts running on a windows server that also accessed active directory. My only exposure to this technology was limited and focused on what I could do with the VM – not with how the VM was created or how I could use PowerShell to deploy a VM.

That professional experience with Azure finally came, and I got to work on PowerShell scripts that could be used to create VMs in Azure. With a script you have something that is consistent and less prone to human error, but even with that sometimes something is set up wrong that can’t be changed once the VM has been provisioned. Alternatively, maybe you are only using Azure for the testing of an application of project, you are done and the VM is no longer needed. You could simply decommission the VM, but if there is no need for the VM to still exist you probably want to delete it and all the parts associated with it.

The first one of these happened to me – there were three VMs where the configuration wasn’t right and it wasn’t something that could be changed from the portal. The VMs would have to be deleted and re-provisioned. I modified the creation script to have the right configuration for these VMs and got those created so work could continue. At this point I started to work on a script that would remove the VMs base on the name and resource group, including removal of all the parts that were tied to the VMs, such as the virtual network and the storage account.

Fast forward to today – I am working on my session for PASS Summit 2017. Instead of creating every part separately with the PowerShell cmdlets, PowerShell is used to apply a template that contains all the parameters and variables for a VM. If you need to make a change it is made to the JSON template, not to the PowerShell. Since I am testing, I am creating VMs and then I want to blow them away right after. I attempted to use the PowerShell script I had used before to remove VMs but that failed – enough time had passed that the cmdlets have been updated and what I ran before was no longer valid. I tried to revise the existing script and then ran across the OMG-I-can’t-believe-I-didn’t-do-this-before answer – just delete the resource group.

If you have contained all parts of the VM in the same resource group, and there is nothing else in the resource group other than the VM, cleanup is this simple. From what I recall about the server migration, there were multiple VMs associated with the resource group. If the resource group had only contained the VMs I wanted to blow away, I could have potentially run this for that clean-up operation. In this case, I was also removing disks from each of the VMs ahead of removing the VMs and had to run a script for that first. Looking back, if everything had been in one resource group, and had there been nothing else in there, I could have run this one line and been done.

For now, this solution works for me, but I know a more specific cleanup script may be more useful – one that will remove a VM and all of its parts based on the VM and the resource group. This will have to happen when I am done with my other testing, but in the meantime, this one line of code allows me to create and blow away VMs while using my personal Azure account…and not run up a ton of charges. Let me know if you find this useful or if you have any suggestions here – would love to have feedback!

PASS Board Elections – Voting Deadline is TODAY!

In case you have been under a rock, or buried in work, or otherwise occupied, the PASS Board elections are going on and end TODAY! There are some amazing candidates (thanks to the selection committee for all of their work in bringing us this fine list of folks) running for a handful of seats. Big question – have you voted yet?

When you vote you are exercising your voice and opinion on who from the SQL community best represents your views on the direction and future of the PASS organization. That said, the total number of people voting during each election cycle has steadily declined, with fewer that 1000 people voting last year.

With this being one of the easiest ways to be involved with PASS, you would think that these numbers would be increasing. If you have not voted yet, go check out the candidates. If you look at all of them, and just can’t decide who to vote for, give that Chris Hyde a second look – while it might seem like personal bias, I honestly feel he is well equipped for this position and will represent PASS and the SQL community with the same enthusiasm and vigor he has brought to the Albuquerque user group and all the other locations where he has spoken for user groups or SQL Saturday events (including the North Texas SQL Server User Group).

The voting deadline is 12 noon PST today (2pm CST). If you haven’t voted yet then put that query down and get to it!