On one hand, it's disappointing that I only seem to update this blog lately due to very significant events. On the other hand, the fact that the two most recent significant events were SQL Saturdays, is awesome. :-) I'm going to use the same format as last time, since it worked well for me.
Morning
To say that my morning was long would be an understatement of epic proportions. The night before, I had been out grooving to my brother in law's Rush cover band, 2112. (Which, by the way, is the best one I've heard in quite some time.) I didn't actually get to sleep until about midnight or so, with an alarm set for 2:30. I figured I've been running on the Very Little Sleep program for some time now, so it shouldn't make much of a difference, and I was right. All I needed was a quick shower and shave to wake me up, and I was on the road by 3:15 AM. The only difficulty I had during the trip was at a travel stop outside of South Bend, Indiana, where the cup carrying my sweet tea disintegrated in hand. Tea everywhere. Outside of that, and my 'slot car' style driving on the Chicago Skyway, which was actually kind of scaryfun, the trip was uneventful. One advantage I had forgotten about was the time zone change. I had timed my departure so that I would be just on time for the first session, unfortunately missing most of registration and the opening spiel. Luckily, I was just on time for registration, so I was happy with that. I got to briefly see a familiar face or two, and meet a few new ones in real life for the first time, further reinforcing the idea that I tweet with some good people. :-) Speaking of which, I should really update my link list over there...
Session 1: Policy Based Management with PowerShell. - Allen White (@sqlrunr)
I know just enough about PowerShell to be aware of how little I know about PowerShell. I decided to keep my mouth shut when Allen asked if anyone needed the 'intro to PowerShell' portion of his presentation. I can research that stuff on my own, on my own time. What I really wanted to get to was how this related to PBM. I wouldn't call what I have here a very large environment, but there's enough that a centralized management system would be very helpful. I'd been looking at the EPM Framework documents lately, and it seemed like this session would be a good supplement to that. I was very pleased with the presentation, both in terms of content, and style. Allen has a good delivery and is obviously knowledgable about his content. p.s.: Allen, if I looked like I was nodding off during your presentation, I will admit I was tired, but I was indeed listening to everything you said. :-)
Major Take-Away: PBM is easily automated - so just do it.
Research Point: Bone up on PowerShell. It's the new black.
Session 2: What to do when it all goes wrong. - Dave Levy (@dave_levy)
We all deal with a fair amount of emergency, outage, and disaster scenarios in our lifetime as technology specialists. Knowing what to do and how to handle things before hand is invaluable for several reasons, not the least of which is looking good (read: experienced). While I was familiar with some of the things Dave described in his presentation, there were a few new wrinkles to me that, as luck would have it, I was able to put to use immediately on Monday morning. (Lucky being a relative position, there...) Most importantly was the idea of proving that an issue was yours to handle, as opposed to proving it wasn't. I saw two examples of that yesterday. In the first, I suspected that the issue was elsewhere, but I went about troubleshooting it as mine anyway. It wasn't my issue to fix, but we resolved things quickly, and as a side bonus, I was able to clean up a couple configurations that I wanted to, anyway. In the second case, while I was pretty sure I could troubleshoot things effectively, I was directly told that, "This isn't a database issue - don't spend any time on it, because we need you doing other things." OK, but with my mad Profiler skillz, I could have saved you about 6 hours of troubleshooting... Turns out they were right, but working together, I think we could have resolved it faster.
Major Take-Away: Don't be afraid to own an issue. If it's yours, you look responsible. If not, you look helpful. No downsides to that.
Research Point: Get with my boss and come up with a work plan for future outages / issues, so we're not just running around theorizing.
Lunch: I went to the car and napped. Major Take-Away: Driving really is tiring.
Session 3: Index Black Ops - Jason Strate (@stratesql)
Indexes are , in my opinion, one of the most inportant things for a DBA to understand. So I was pleased as punch to be at Jason's talk. Really what it was, was a thorough examination of the DMF sys.dm_db_index_operational_stats. I do like my talks on the technical side, and the advantage of something like this, is that I can begin putting it to use first thing on Monday morning. This was the only presentation, though, I had some issues with. I think Jason will admit that the demos he had didn't go well, and I would say he handled the difficulties well, but not the same way I would have handled them. I've done a bit of presenting, and if I have something in a presentation that doesn't go right, I don't spend time fixing it. I suppose it may be different when presenting SQL than when presenting Skepticism topics, though I would think each audience can be equally unforgiving. Both audiences have people with laptops fact checking you during your presentation, I suppose. Still, I would call less attention to the gaffes. Despite all that, this was yet another excellent presentation and gave me more material to work with.
Major Take-Away: Need to add the sys.dm_db_index_operational_stats to my stats gathering scripts, to supplement the missing index / index usage stats.
Research Point: What other DMVs/DMFs should I be gathering from?
Session 4: Partitions, Filegroups, and Files - Brent Ozar (@brento)
I think most people will agree that Brent gives an excellent presentation, so I'm only going to say one stylistic thing about his presentation. It was clear that the audience knew their stuff, and it was cool to see Brent get excited about the questions and points the audience was making. A presentation that turns into a dialogue is usually a good one, in my experience. I definitely learned a lot in this session. In fact, I came out of this presentation with a bit of a laundry list of things to do, in addition to Brent's checklist of "things to do on Monday morning." Easily my favorite presentation of the day.
Major Take-Away: Don't cross the streams. Er... I mean, don't separate indexes and base tables into file groups - no inherent advantage. (You listening, boss?)
Research Point: Find out what the critical tables are to our main application and get those into a quickly restorable file group.
Session 5: SQL Internals, It's what's for dinner! - Christina Leo (@christinaleo)
The only presentation that made me want to throw my laptop. I will explain in a moment. I've been to a couple of internals presentations before, and was roughly familiar with how they work, but like other classes and courses of study I've engaged in, any amount of reinforcement is a good. I really appreciated the work and originality Christina put into her analogy for the presentation, which was the art of making dinner. Each step in the process of making a meal corresponded nicely to the process a query goes through to return results. Her explanations were clear, concise, and seemed fairly well polished. Given that she was a first time presenter, I would have expected a bit of nerves, but once she got going, everything ran pretty smoothly. The information was useful, and I enjoyed that session very much.
Well, almost. And here's the only thing I didn't like about this particular SQL Saturday. I've only been to two others, but this was the first one where I've seen participants be actively rude and argumentative towards the presenters. Not sure if anyone else saw much of that, but in at least two presentations I was in, including Christina's, there was at least one person who I wanted to chuck a heavy object at. To the gentleman in the rear of the room, who chose to argue about ACID properties: in the first place, you're a jerk. In the second place, I checked, and you are WRONG. I must, at this point, thank Christina for handling herself very well despite the rude interruption. I'm not sure I would have been as graceful in her position.
Major Take-Away: Need to gather more performance metrics on my log files / buffer cache.
Research Point: What's the momentum of a Cr-48 netbook at about 45 feet/s?
Dinner and Karaoke:
After all that, I was famished, so IHOP provided ample fuel for me to continue my day. I think I might have seen Shelly Noll (@shellynoll) around there as well, but she wasn't sat in my section, so I pretty much had time to just sit and continue my note-taking. I took a lot of notes. :-)
After dinner, was SQLKaraoke, which is always a favorite activity of mine. I rehashed a couple of my favorites, (Pearl Jam's 'Even Flow' and Hoobastank's 'Crawling in the Dark'.), but finally got to hear Wendy Pastrick (@wendy_dance) sing, and was thoroughly entertained by everyone else. The singing ability varied widely, but as I've said before - the ability to sing and your ability to do karaoke have nothing to do with each other. The best karaoke performers I know can't sing a lick, and don't even really try to. They just get up there and have fun with it, and I could literally listen to them all night. (And have done, I assure you.) I only got a couple of pictures, but really enjoyed myself. I was sorry to have to leave early, but the drive home was a long one, and I had no plans to stay in Chicago that evening. So 13 hours in Chicago, and 12 hours on the road. Yeah, I'd call that a full day, but worth every minute. Looking forward to doing it again next year.
The list of people to thank is endless. Everyone mentioned above, plus all the volunteers and sponsors who made the event happen.
See you in Columbus, at SQL Saturday #75? :-)
Thanks.
-David.
No comments:
Post a Comment