Much of the IT world is filled with the facts, with what we know. If I do “x”, I get “y”. But that’s not true for everything in the IT world. Especially not the DBA’s world. Therein lies several areas that feel like they’re shrouded in mystery and only a few know the magic words to fix the problem. One of those areas is performance tuning. The problem for the DBA, whether seasoned professional, newly voluntold, accidental DBA, or the mid-level DBA striving for ever increasing knowledge and skills, is where to find the magic words.
Sure, the DBA can turn to the web. Blogs and forums offer manifold solutions that may contain helpful information, many don’t. Some even offer harmful information. One of the most frustrating things about the web is that the information is like a 10,000 piece jigsaw puzzle with only a rough pencil sketch on the box lid of what the puzzle should look like. As a DBA, you know a piece about indexes, a piece about tempdb. You know enough to understand your transaction log on a certain database is highly fragmented but not enough to be entirely certain on how to fix it without affecting the other databases on that server, which really aren’t too bad. To make matters worse, your environment isn’t homogenious (I mean, is anyone’s?). You have SQL Server versions from 2000 (the ones everyone said they would upgrade before finding out that they couldn’t) to 2014. So now you have problems on both the lower versions and the higher versions. You ask yourself…. “What is a DBA to do?” Well, the answer to that question is to take Paul Randall’s and Kimberly Tripps IEPTO1 entitled, “Immersion Event on Performance Tuning and Optimization – Part 1” in Chicago October 2-6, 2017 where the the mysterious is made mundane and the pieces of the puzzles, even the pieces that were lost, are found and put together. And the picture suddenly pops. It’s not just a pencil sketch but gloriously colored like a Bing screen saver.
If you haven’t guessed, I’m one of those voluntold DBAs. Before I was a DBA, I was a programmer who was on a team of programmers assigned to choose a database for the company. Weird, huh? Programmers choosing a database? But, that’s what happened and that was when I was voluntold. I then became a database developer and a DBA….all with the same order from on high. Well, it’s been a while since those glorious days when I was absolutely terrified to touch the keyboard, knowing I had to do something on a live server. Did I research? Heck, yeah! But, I was still nervous, worried that something would go wrong that I couldn’t fix. Since that time, I have worked hard to increase my knowledge and skill. I’ve done well. I’ve progressed. But the one thing I haven’t been able to master, so far, is performance tuning.
I have some fantastic SR DBA’s to work with. Man, I love these guys! Not only are they a seriously deep well of knowledge, they pour that knowledge out all day long 7 days a week. They laugh and tell me that I know more than I think I do. But, to be honest, one of the areas that causes me angst is performance tuning. They laugh (kindly) at me about that, too. Again, telling me I know more than I think I do. But here’s the kicker, remember those puzzle pieces I told you about? I don’t have all the pieces. Worse yet, I know there are some missing, but I want quality information to make the best decision, not only for me but for my company. I want to excel in my knowledge and skill. Like my mom always says… knowledge is power. That’s what I think Paul Randall’s and Kimberly Tripp’s class is going to give me. Knowledge. Which makes a difference.
I just transitioned to a new job from a hosting company. Dude, you think you have a big environment! The hosting company had 30,000+ (yes, you read that right) databases, 300+ servers in 6 data centers around the world. The smallest database was only a few 100 megabytes. The largest was over 10 terabytes. The problems were endless. Blocking was the serious issue of the day. And have you ever thought about the issue of backing 30,000 databases up in only 56 hours over a weekend? Or had multiple big clients say, all on the same day, “Hey! Our queries aren’t finishing? What’s up with that? Can you help us?” Or how to complete indexing on 800 databases on one server in a few hours over night. OLA doesn’t cut it in that environment.
My new company doesn’t have that many database or that many servers. And, remarkably, all their database servers ARE the same version and service pack level. However, what they do have is a highly transactional database with millions of transaction per hour. Right now, they’re working on some really cool stuff which is the future of the company. The transaction count is expected to be even more through the roof when these new products are released. I want to have the knowledge to help scale these servers out and bring that knowledge to the table with confidence to help my company grow.
So, what are my nemesises for performance tuning? The things that make me loose sleep? Blocking and scalability.
I feel certain that Paul Randall and Kimberly Tripp while reveal all … okay, maybe not all but a lot of the myserteries of SQL Server performance tuning. And I would love the opportunity to learn.
Class linked and referenced: IEPTO1: Performance Tuning and Optimization – Part 1 class in Chicago, October 2-6, 2017 https://www.sqlskills.com/sql-server-training/chicago-iepto1-20171002/