I just got my comments and scores back from SQLbits 8 for the two sessions I did during the conference itself (I know how the precon went, and that was great). Overall I’m fairly happy with my scores, especially given the fact that I was really sick that entire trip. Over the course of three days I delivered close to 12 or so hours of content with a really bad ear infection that made my head feel like it was swimming. Not exactly ideal, and definitely not recommended.

I’m not going to give scores, but address some comments made. 

My first session on Friday (“So Many High Availability Options, So Little Time”) was in a room that was being partitioned after the keynote and due to the changeover, started very late. Most of the numbers were 7, 8, or 9 (out of 10). However, what I do wish people would do when they dislike your talk (or aspects of it), give some constructive feedback on WHY you don’t like it. I know I’m not perfect, but I can’t improve if you don’t give me any kind of constructive feedback.

My second session on Saturday (“Multi-Site Failover Clusters with SQL Server 2008 and Denali”) overall went really well, people seemed to like it, but a few felt that the expectations were not met … but yet failed to tell me why in most cases. I’ll take that ding because clearly I missed the boat somewhere and probably in my description. My guess is that those who were dissatisfied were for this reason, which someone wrote “I was expecting more details about Always-On option.”

While a fair comment, here’s my session description:
“Planning a failover clustering implementation in a single data center is fairly straightforward, but what happens when you want to span data centers? While most of the considerations are the same, introducing distance complicates matters. The story is also more complex with SQL Server 2008 and 2008 R2 than it will be with Denali. This session will cover how to plan and deploy a multi-site SQL Server 2008 or Denali failover cluster using Windows Server 2008 R2.”

Nowhere in my description do I talk about AlwaysOn Availability Groups, which is the new feature of SQL Server Denali. So I’m not sure how I didn’t meet an expectation that I didn’t set. I guess throwing Denali in the title and abstract people assumed I’d talk about AGs. I didn’t really address it much in the session except briefly since it wasn’t what core I was talking about. I also have to say this: there was not much that can be said about Denali publicly past CTP1. So even if I could talk about AlwaysOn AGs, I couldn’t demo too much or talk to things that were not in CTP1. That’s one reason I didn’t really do much about  AGs in any session.  So I apologize if you thought I’d do a deep dive on Denali AlwaysOn AGs. That was never my intent.

I do want to address this specific comment: “Probably inadvertently, but by requiring the attendants to ask questions later, the presenter discouraged the audience to seek clarifications and/or contributing with it own challenges/opinions, which I think is a significant part of the value we get from these sessions.” (And for the record, the person did give me good scores.)

I think this could be a difference between the US and Europe or the UK. In a pretty small session (training, under 10 people  where it’s really intimate) I sometimes don’t mind taking a question or two during a session, but when you’ve only got 60 or 75 minutes including Q&A to do your deck and demos, a talk can get easily sidetracked or derailed if you allow questions during the main part of the session because often one question begs another. Asking people to ask questions during a formal Q&A at the end really isn’t strange at all. I’m sorry it may have inhibited this person or others, but the thing I absolutely hate to do which cheats everyone is to start taking questions and the next thing you know, you’ve got 10 minutes left to do your slides and finish your demos. As a speaker you run the risk of comments like this. This is a ding I’ll happily accept. I’ve had a few sessions go to the wire where I didn’t get it all in. It’s bad for you and it’s bad for me.

Here’s another: “Allan is excellent, would have perhaps liked a little more demo but overall a valuable session.” I know the commenter, but let me ask this: how many of you are schlepping variations of multi-site clusters around? Go ahead. I’ll wait. <insert Jeopardy theme here>

Needless to say, not many people have the hardware to set up a demo multi-site cluster back home let alone on a laptop (and I had two setups under Hyper-V – SQL 2K8 R2 and Denali CTP1). I’m not sure what else I could have shown. I know it would have been more impressive to roll in a bunch of hardware, but that just ain’t gonna happen. I would be curious to hear from this person what he thinks I could have shown that I didn’t.

This comment I can address a bit more: “Was expecting more demos and details on Denali rather than 2008R2” Like I said above, Denali at the time was (and still is as of the time of the writing of the blog) still at CTP1 for folks to download and try. Even if some other build was available, it would not be something that could be demonstrated publicly without the approval of Microsoft. Because the UI may change, I deliberately did not show the Setup screens in CTP1 for configuring a Denali multi-site cluster. I did talk about what you need to do at the Windows level, which in my opinion, is really the important stuff. You need to get Windows working. SQL is easier in comparison in this case because most customers today do not span data centers for clusters. Since current tech is what most people will implement for the forseeable future, I did focus a bit more on that BUT nearly everything from SQL Server 2008/R2 is applicable (with some differences that I pointed out specifically at the Windows failover cluster level) to Denali. At its core, they are very similar in that way. I didn’t think I needed to bop people over the head with that lesson. Maybe I need to be more explicit.

Again, I’ll ask the same question of this other commenter: what demos would you like to have seen with regards to Denali that I didn’t show? Once the instance is installed and clustered, it’s just a clustered instance of SQL be it SQL Serer 2008 R2 or Denali (and one of the things I made sure to point out were differences between SQL 2K8/R2 and Denali as it will appear in Failover Cluster Manager).

This one is a bit hard for me: “I was expecting a lot more from this session. Having read Allan’s book on SQL2005 HA. I was a big fan of Allan’s, I think I was probably expecting a lot more and was a bit gutted as I had to decide between Allan and Klaus on the training day. I decided to go to Klaus’s training day session. So it may be that I was bitter that could not attend Allan’s training day session. Thats my fault. It didnt help that Allan was a bit under the weather”

I hate to disappoint anyone. I’m human. I know I’ve been to concerts where I walked in hoping for the most amazing experience and then at the end feeling underwhelmed. I’m sorry that happened to you – and that’s sincere. I woul dbe curious to know (as with the others) what more you were expecting. Clearly I missed something if it came up more than once.

On the flip side, I love these

• “I will be telling my boss and directors about this session – it was that important”. Awesome! THAT’S why I do what I do.
• “One of the best presentations of the conference for me, highlighted a lot of the pitfalls and what needs to be considered as well as the expected features for the next release of SQL. Has also had a profound impact on my career path. Excellent, excellent session.” Wow. I’m blown away – I don’t set out to impact someone’s career path, but just wow.
• “Nice guy who knows his stuff. Would like to see more about clustering at future SQLBits events.” Hope to get back there! Definitely some interesting stuff coming in Denali for HA.
• “Allan was another great speaker who I’m sure could speak at length for many complete days about his experience. After spending the Wednesday in the HA pre-con, there was still more to learn in this session. Very approachable and
  knowledgeable.” Thanks! I tried to minimize overlap.

So that was my SQLbits feedback in a nutshell. I’ve got six months between now and PASS (assuming I get a session) to come up with some real whiz-bang Denali demos. Better get cracking!

I also hope to get back to SQLbits in the future and right some of these wrongs.