September 2007 - Posts
I almost forgot to record the podcast for tomorrow's editorial. In case you're a glutton for punishment, it's
here. I'd spent over an hour polishing off the writing part of this in the am, not having finished it on Friday.
So by fast I mean it took only about 10 takes. After about 4, I decided to break it up and record the intro (title, name, etc.) and separate as well as the credit for Incompetech. BTW, if you like the music, check out some other stuff from Kevin. I emailed him after sampling a dozen or so tunes for permission. I think he's done a nice job and it's a good set of music since I can't get licensed for commercial stuff from the radio. At least not yet.
After bathing the kids and getting them in bed, I was just about to approve the send of the newsletter. I'm behind with the new system, so I haven't quite gotten used to scheduling too far ahead yet. It was then I realized that I needed to record, so I ran downstairs and fired up the audio programs. You might get a kick out of the mistakes I make while recording, but it's a bunch of work to edit things together. But I said it was Friday, stumbled over the title, got tangled up in the second sentence, said JD Peoplesoft, and more.
I'll save some more bloopers when I do video, it will be more fun.
Though it kind of looks like that. Here's the first video podcast.
Halo 3Actually it's probably the 20th, but it's close enough for a test, beta release. Hey, this is a Microsoft site, so we'll release stuff that isn't all the way there. Heck, I don' think it's any worse than the stuff I've seen a Lockergnome, and that's popular. Lighting is an issue and until I figure out a good routine for moving stuff off the little camera, I thought I'd try the webcam a bit.
This certainly gives me more respect for the people out there producing them, like GeekBrief. However they've spent about $10k on hardware, which I can't really justify right now. Maybe if we build more popularity, but for now I'll work on improving things here. Maybe some old lamps in the office will help.
I responded in an interesting thread this morning where someone was looking for advice on what to monitor on their SQL Servers. As expected, I can't find the thread right now and I'm rather busy, but suffice it to say someone was wondering what to monitor on their SQL Servers using some third party tool.
This was interesting because I had actually seen a bunch of Kevin Kline's presentation on benchmarking at the recent PASS 2007 Summit. I like Kevin and he's got great information, but I was surprised at how many counters and areas he was looking at benchmarking. If you have a few critical applications I guess it makes sense to gather that stuff and I'm sure it's invaluable when troubleshooting, but I've rarely had a tool to handle that many counters. Especially when I deal with so many servers. If I was a 1-2 server guy, which I guess I am now, I'd look at a lot of counters and when a user complained, I could check all kinds of stuff.
But in larger environments where I've worked, with 10-100 servers, it's too much to track. And it's too much to look at. Personally I like the rough thumbnail approach. I tend to look at the rough performance of a server. CPU, memory reads, disk I/O, users, and transactions. Those items are a good gauge of how your server is doing overall.
However the numbers are meaningless in a real-time environment. If I tell you your server is running at 80% CPU or at 500 memory reads/sec, is that bad?
You can't tell me. If you average 85% CPU and 400 reads/sec, then that's pretty good. If you average 30% CPU, that's bad. So benchmarking is really important to understanding where you might be slow when someone reports an issue.
But do you need them to report an issue? Can't you just set an alert for 85% CPU and check the server when it comes? NO!
At least I won't. There are so many variables the server responds to and ends up having it's performance jump all over. Ever watched your counters when a log backup or db backup runs? They might spike all over and you don't want to get alerts everytime that happens. You can ignore them, but then you might tend to ignore something that's really wrong as well. So I've avoided setting alerts except for when we have real problems and I'm monitoring on a regular basis. They don't last long, usually just days.
This is one area that Microsoft would do well to build into the product. Or make it a separate product that's REASONABLY priced. Note: Most of the products, Patrol, Unicenter, etc. that do this cost way too much.
Instead, give me a simple traceing setup that will capture my data and let me graph it. The Health and History tool was a good start, but it was half-baked. When the timer pops out, send it along.
It's an interesting question that we had come up, and honestly I'm not completely sure what I want. I'll describe it a bit here and then you can send comments to me (sjones at sqlservercentral.com) or
post them in the poll I've set up.
Right now we have each forum post equaling 1 point. Each question from the Question of the Day has a point value (which is randomly chosen by me, usually defaulting to 2) and you get the appropriate number of points for each correct answer. So your total is
Points = Posts + (correct questions * points for each correct question)
This has confused people and while it's an interesting idea, a bunch of people, including myself would like to see their post count. We could easily show posts in the forums and points elsewhere, but here's the trick.
Right now member levels are stored in one place. So if we separate things out, you could have one member level in the QOD section and another one in the forums. Changing that is a little tricky, takes some development time, and more, we need to think about how we'd like that to work.
Personally I don't know if it matters if you have different member levels, and while I'm sure some people won't like it, I wonder if they're the majority. Let us know.
That is for sure.
We had a development meeting today with my publishing team (meaning I'm on the team, not leading it) and the developers. We went over some of the bugs and issues we're having. Overall it's not too many bugs, but mostly issues with changes in workslow and people's patterns of use of the site. Of those, we decided on some high priorities, some lessor ones to work if time frees up and then some that we will hold off on and evaluate. The plan is for a weekly development/publishing meeting on Wed and regular updates (meaning daily) from the developers so we can try and get things moving.
The high priority items for now are:
- Get email notifications for replies to default to ON
- Separate search into sections (scripts, articles, forums)
- Move forums off the front page to a child
- Codezone - We've dropped off and want to get back on for Books Online searches.
There are a few smaller items and some that will be simple text changes, but that's the high list for this week. Changes to briefcase, QOD, etc. are slated next on the priority list.
There's a little bit of conflict between my group and the other Red Gate site. There are changes that need to be made to other sites and so I'm competing for the developers' time. It's a big of an argument between my boss and theirs, and we'll see what happens. SSC is important, but in the grand scheme of things, it's not as important as software.
I realized today that I'm not in charge of this project. I have a lot of input, but I'm really not the guy driving the train. That's ok and I can deal with it, but I was thinking I didn't have to manage things, but could really then force people to bend to my will. Not the case and it was a good realization for me. I also have to keep in mind that I'm one part of a publishing group and a small part of Red Gate. We don't pay for ourselves directly, so we bend a little to the whims of other groups at times.
However we are an important brand and so we do get to call on some resources. Search looks like it might be the big issue here as it's hard to use, but we'll see what happens.
As with any project, despite being fairly pleased with the final product, there were bugs. In fact, quite a few issues that cropped up,which we hadn't seen in testing. Of course we could have done more testing, but part of that is knowing what to test and part of that is understanding usage patterns.
This is a community site, so we try to balance the needs of different groups that use the site to come up with something that works. There were a few design decisions that our developers chose that didn't fit the site well. Despite beta testing, we just don't get the same usage that a person would get in normal usage and the complete redesign of the database means that we also had issues running this as a live site in line with the old one. In retrospect, it would have been cool to have some merge replication set up to move forum posts and scripts between the two sites and run them side by side, but we decided against that and so here we are.
The biggest item that I can see right now is the auto-subscribe for posts that you reply to. In the old system this was a per-user checkbox in your profile. In the new system, it's gone. So we have a sitewide setting, or at least a default. I haven't found a setting, but I've been working at a somewhat frantic pace over the last day trying to respond to issues and forward things on to the developers.
The whole process shows how unorganized we are in this mode of development. Most of the Red Gate development takes place over much longer timeframes. And point releases are built on a schedule of weeks or months. We're trying to work in days, and while we've deployed a few fixes, I'm concerned over moving too fast and breaking things. We also don't want to yo-yo to the whims of a few people.
So I've asked for some coordination. Someone that can both PM this project as we develop a point release and also manage the developers, ensuring they're working on what the community wants done and not what they want done. It's a little slow as we ramp up, especially with the time difference, but we've collected a lot of items in the bug tracker and someone should be estimating their resource load (people and time) so we can have an early morning development meeting and set priorities. Our goal is a Thur or Fri point release this week and I'd hope next week again. I'd like to see weekly releases over the next few weeks as we try to clean things up.
It's done, it's out, and check it out!!!
I guess if you got here, you've probably noticed :)
This was a fairly painless upgrade. The boys in the UK got up early, came into the office and converted over the database, moving the website DNS and then letting me know it was done. By the time I got up around 8 this morning, things were complete and we were testing. Other than a minor permissions issues, things are looking pretty good. I did have to run some T-SQL code to match up new editorials (added as articles) to the existing forum threads, but that's not a big deal. We hope to automate the backfill next week sometime and all new ones work fine.
The biggest changes are on the back end, in the tools I use and how I run the site. However overall I think it makes for a more consistent site that works better. I know that sounds silly, but it makes sense. And it means I'm officially off the code developer team. I have no idea how this AJAX, .NET, Javascript thing works, so all my requests now go into a Q. Hopefully I'll keep the developers, or at least one, for a few months to add a few sections to the site.
Now if we can just get the Google indexing caught up :)
I haven't seen a lot of sessions at the 2007 Summit, but I did go through 4 or 5 across the week and I noticed a few things in all of them. There were some first time presenters, and some nervous ones as well, but even the experienced ones could do with a few tips. I'm not the best speaker, but there are a few things I'd recommend.
- Don't read slides - This is the big one. We're all big people, many of us have the slides, and we don't need you to read them. Don't even read a bullet point if you're talking about it. Just assume we can read it and talk about what it means or the point you're making.
- Speed - If 3 people have asked you questions and you say "I'll get to that later", you're moving too slow. I don't want you to talk fast, but you need to keep up the pace. It's better if we're overwhelmed in 75 minutes than underwhelmed.
- Background - In line with speed, move quickly through background. We're all pros, we do this stuff and we want to hear about your idea. Don't spend too much time setting things up and do it quickly, 5-10 minutes. Practice this and if people are lost, they'll ask questions.
- Focus - By this I mean the presentation should be focused. I saw a few presentations that were so broad, that they didn't really offer practical, sound advice. Too much information was presented, but it wasn't overwhelming, it was over load. Nothing was covered deeply. Focus your presentation on a problem. If you do lessons learned, do one every 10 minutes, mention the lesson and give some reasons behind it.
- Repeat questions - Not just for us in the audience, but the audio recording as well. Repeat what you're asked.
- Relax - We want to like you. We think you've got a good idea or we wouldn't be there, so have confidence and relax.
I was late, not getting up quick enough and then fighting traffic into the city. From there I did a little work before hitting a session that I was looking forward to attending.
And I was completely disappointed. I won't mention the session here for the world to see, but it was horrible. I heard that comment, along with "it sucked", and that it was a "waste of time" in the hallway as we left and I concurred.
It was bad enough to stick two people up there who weren't the best English speakers (both came from overseas), but on top of that they didn't really give any technical information. They told us there was an NDA in place so they couldn't be specific, but then presented a general schema/design in 40 minutes that could have been done in 5. They didn't give specific recommendations or even talk about challenges of working through the area the session was supposed to focus on. Then they blew off some questions that people had asked. I could have read the slides and given a presentation that had a similar amount of information.
It was a 400 level session, so I expect things to move quickly. This was a 100 level session that didn't provide any great information and the slow pace prompted the audience to keep asking questions that were deferred to later slides and then not answered.
Hire some actors and give a good presentation that provides technical data, not specific, but solid technical data about how to handle similar situations. Very disappointing.
The only interesting thing was hearing that partitioned views don't support RI.
After that it was a hectic lunch talking with people and then doing an interview with a PASS board member.
After the PASS opening, which I dropped in long enough to get a beer, we had the SQLServerCentral.com party.
It seemed to start a little slow, but ended up busier and running smoother than any other of our events. I have to credit the CCEvents and Boulder staffing people that are supporting PASS here with doing a great job. We probably had 120-150 people in there, all playing casino games and enjoying a few sweet snacks.
By the end of the night I was worn out. Tony Davis and I were walking around giving away lots fo DVDs and other prizes to winners in various card games throughout the night, but it's just hard to give away 60-70 prizes. Towards 9pm I started just giving them to random people that were walking around since we had a few people that didn't want to play games.
I got to see a lot of friends from previous conferences and meet a bunch more people. It ended up being a really good event for me and without Andy and Brian, forced me to get more involved. I wish I could have sat more and talked with some people, but overall I really enjoyed the event and I hope everyone attending did as well.
However I think I'll go with a few less prizes, maybe some newer DVDs and make it a little slower pace next year.
I'm kind of stunned. I have asked 6 people to go to the PASS conference over the last 4 days in Denver, all local people, and they cannot get time off for the conference. And a few of these are DBAs!!
Two are in projects, granted, so time off is hard, but 2 are DBAs and their boss won't authorize time off and they're frustrated. Needless to say I'm looking to find some new jobs for people as well now.
A conference is a debatable ROI for a company, and I can understand not wanting to shell out the $2k admission, plus $1000 or so in travel/food and the time off. But for local people, all they need is time off. No food, not airfare, no admission. It's an easy way to get some training, retain an employee and maybe have them get some ideas or solve a problem. After all the SQL developers from MS are here and they're great about answering questions.
So I got a final count of referrals from the PASS conference yesterday and it was 183, so thanks to everyone that signed up!
We get $50 for each one and my job is to blow every bit of it. This year I was a little worried with the conference coming earlier and a slower signup than expected. I almost didn't order Best of SSC.com books to not mess up the budget. Now I hope the books get here.
So to date I'd spent about $2k on prizes, $2k on the casino company hosting the event, $3k on shirts, $1.5k on books, so about $8.5k total. In doing the budget today I had $1100 (roughly) left to spend. So after lunch I headed to Best Buy and started wandering around. My goal was a couple big prizes, but of course they didn't have a few things that I saw on the web and I changed my mind.
This is one of the best parts of my job. The best being giving things out on Tuesday night. I pick out prizes, this time updating my budget in real time on my T-Mobile Dash. I found a few bargains, like a 320GB external drive for $100 and a GPS for $179. I picked up a few smaller prizes to up my count. I didn't get enough for half the people, but I think I'm above 1 for every 3 people.
I got some Scene It's on sale, a couple "Deal or No Deal" games, some walkie-talkies, some mini-arcade games, a portable DVD player, a few more iPods and who knows what else. The guys at Best Buy think it's a riot and there's one guy that's seen me for 3 years and just laughs when I come up with a cart full of stuff. They all want to know where they can sign up to get some stuff.
I got to the register and the total was $1095 for my stuff. Right at budget.
So I treated myself to a Java Chip from Starbucks (that's my "handling charge") on the way home, getting me back with $1099 spent of my $1100 budget.
There might be a few more prizes, depending on what I hear from PASS on Monday or Tues.
Editorial coming next week, but so far
this is amazing to me. Microsoft's stealth updates of their client update tool.
Typical explanation from the PM, "We think it's necessary so you can get updated". It's the arrogance that we know what's best and the disbelief that we wouldn't agree with them.
I know you're smarter than me. That doesn't make you right.
OK, here's
one take. Not sure if it's any good, but it's a start.
No live editorials yet. I've been working through this process and trying to figure out what works. If I can get a good take, maybe I'll put tomorrow's out there.
I'm working on getting a small studio set up for some podcasting of the editorials. That means I put a mike on my desk along with a pop filter. I tried to do some basic recording, but the problem so far is that the levels are too low. I tried the free "Audicity" download and it works pretty well, allowing me to combine multiple tracks into an MP3, but the volume on the mike still isn't quite loud enough.
I went and bought a pre-amp today, getting talked into a USB model from M-Audio, but it doesn't seem to work, or at least I can't get it to. I should have stuck with the basic mixer idea that I had and gone with that.
Actually we might do some video as well and once I get a good pipeline of editorials ready, I'll start releasing them.