February 2011

February 23, 2011

The Man (or Woman) Behind the Curtain

It’s amazing how connected we all are today with things like Twitter and Facebook. The SQL Server community is pretty active on Twitter (you can follow me, too), and a little controversy was sparked which I wanted to comment on because I’ve been a victim of similar circumstances in the past.

Peter Adams (blog | Twitter) wrote what he probably thought was a simple blog post callling out Paul Randal (blog | Twitter) on being wrong in in a tweet. I don’t believe he intended to stir up any controversy, but boy did it spark some healthy debate in the SQL community.

Let me say from experience that Twitter is not the easiest medium to get a complete thought out in 140 characters, and responses are generally off the cuff. Paul even couched his tweet by adding “I believe so …” and owned up when he was wrong.

Now, I don’t think Paul went and crawled into a fetal position when he found out that he was incorrect. People you personally consider to be experts CAN and DO get things wrong from time to time. Do hitters in baseball bat 1.000? No! The best hitters are generally 1 for 3 (give or take), and SQL experts have a much higher rate than that. But they do occasionally have a swing-for-the-fences strike three.I can say from experience that Paul is one of the sharpest guys around on SQL Server in general, and especially the relational engine. That isn’t only because I know Paul, but if you look at what he has done over the years and his contributions to the community, it’s really hard to argue that.

I know it can be a let down when you find out someone you admire is human. I can remember when I first interviewed Dennis DeYoung back in 1996 and then met the band prior to a concert where I took photos for the Keyboard Magazine article I subsequently wrote. (Full disclaimer: if you haven’t picked up on it in my books about how I like the band Styx, well, you know now) Growing up, I thought those guys were untouchable and perfect. You have an almost mythological view of what goes on in a band. I can still hear the anger in his voice when I told him about a promotional CD sampler that their record company at the time (A&M) released that had songs that shouldn’t have been released. Dennis is a nice guy who I’ve always gotten along with, but that conversation got very serious, very quickly. He was NOT happy. That took me a bit by surprise and gave me a bit of a glimpse behind the mirror. The final shattering of the mirror came when I went to take the pictures. I went backstage prior to the Hartford, CT, show and before they did soundcheck. Dennis was nice to me, so was Chuck Panozzo (the bass player) when I happened to run into him (he’s actually an influence on my bass playing, even though he’s not Mr. Flashy). I didn’t see Tommy Shaw or Todd Sucherman but I can remember running into James ‘JY’ Young. He was sitting on a chair and when I approached him to just say hello, he was very cool and aloof. That put me off. (When I interviewed JY a few years later for the album Cyclorama for my alter ego, StyxCollector, it was just as weird … but that’s a story for another day.)

I realized these guys were human just like you and me. They get angry, they sometimes don’t want to deal, they put their pants on one leg at a time … well, you get the idea. I walked through the mirror. Since then, I’ve interacted with members of the band and it’s been great for the most part. I just talk to them as I would anyone else. I don’t put them up on a pedestal and I think they like the fact I don’t come at them as a fanboy.

I wish the story ended here for Paul and Peter. Sadly, it doesn’t. Full disclaimer mode: I’m not here to vilify Peter. After all, he did just tweet this: “Comments approved. I cherish freedom of speech ;)”. All I know of him is the blog posts and Tweets I’ve seen. I’m sure he’s a smart, stand up guy with good SQL Server skills, has a great family and friends, etc. He seems to want a MCM for SQL Server; you’re not some entry level DBA if you want that.

After that first post, he then posts a blog entitled “MCM videos and repeat of it in the SQLskills Immersion Events “.

Here’s the deal: everyone learns differently. Some prefer to sit in a class, others can just watch videos. When I deliver my clustering class, I’ve given parts of it at things like PASS, but unlike Paul, Kim, and Brent, I didn’t basically put the whole class online as they have. That is a ballsy move.

Those MCM videos are a wonderful resource for the entire community. I’ve watched a few and even for those of us who have been around the block and know a thing or two, they are a great refresher on some nuances of the basics we may have lost over the sands of time when we’re doing things like cramming clustering in our heads. AND THEY’RE FREE (go watch! now!).

The one thing a class affords you the opportunity to do is to ask questions in person, whiteboard, and interact with others. I know when I teach I also learn from my students. The best teachers are also great students. The best classes are interactive ones. When everyone is just staring at me and not participating or asking questions (PASS presos aside where you just take them at the end), I know it’s going to be a long day or week. Going to a class – even for material you can watch videos online for – can make the content somehow come alive. Hopefully you’ve all had moments like that.

Paul tweeted about how no class is 400 level all the time. It’s so true. Whether it is 2 students or 200, people will be at all different levels. If you assume everyone is at 400, you’ll lose a good portion of the audience. A good course writer will start at 200 and get deeper. It’s not a bad thing to reinforce some basic concepts even if it’s done quickly to ensure everyone is on the same page. It’s a basic level set. I do it, and I occasionally get the SAME ding that Paul is complaining about.

“I know this already” – yeah, but does your classmate? Have I been bored in classes or sessions at times? Sure, I’d be lying if I wasn’t. But once you get past what you know (or think you know), you hopefully pick up some tips that are invaluable, or learn different approaches to things you do. You need to go in with an open mind. And let’s face it, even 1 or 2 days of all 400 level content is too dense for most folks. Digesting deep content sometimes takes time to sink in. The right instructors can teach novices who don’t know anything as well as field the weird scenarios from the people who want that deep level of content. I hope I can count myself in that camp with some of you.

Does it end there? No. There’s a final blog post, “Microsoft Certification Trash Talk” where Peter posts some of Paul’s tweets about his thoughts on certification. You know what? I 100% agree with Paul and I’ve said it for well over 10 years. I can remember a conversation around 2001 or 2002 I had with a customer. They had a process that was taking on the order of like 14 hours to complete on SQL Server 7. By doing some optimizations with indexing and such, I got them down to under 4 hours I believe (I forget the exact number). I only had a little bit of time onsite, and I’m sure there were other things I could have tweaked. They were psyched. At the time I believe the cert was the MCITP/DBA one. The company asked me, “Can a certified MS person do what you just did?” My response was, “Only if they have the experience.” They were taken aback by that.

I know to get past HR who is screening resumes a certification has become de rigeur. The problem is, without the experience – at least to me – it is somewhat meaningless. In my mind, a certification should be something you do that cements your experience, not substitutes for it. Having taught some modules during two different rounds of MCM when it was a three week endeavor, and devised some training material, I can tell you that MCM is not your average certification. Kudos to Joe Sack and his team for coming up with a way to scale it, keep it real world, AND cheaper! Ask anyone who has taken it and passed if it was easy. I’ll bet they’ll tell you that it wasn’t, and they are some of the smartest people I’ve encountered on SQL Server.

If I was heading up a DBA staff, I would hire the people with the right experience. Would a certification be nice on that experience sundae? Sure, but I’ll take the guy or gal who has 10 years of experience in the trenches who can give me war stories about how he (or she) solved this problem or made it though some horrible crisis versus someone who has barely been around the block but has a lot of certification letters on his resume. Other hiring managers out there may see it the other way around. Of course there still needs to be the soft skills to go along with the DBA skills, but that’s not the point here.

You can’t teach experience. You can learn for a test (well, most tests anyway).

See the difference?

February 15, 2011

Tweet, Message Board Post, Blog Post, Whitepaper, Or Book?

As I’ve started to use social media a bit more, it’s been an interesting experiment. Those who know me know that before blogging, Facebook, and Twitter, I did just a little bit of writing (insert sarcastic laugh here). I resisted Twitter for awhile and am still a bit iffy about it to be dead honest. I’m warming up to it.

Using Twitter has been informative on a few levels. Twitter has a basic limitation of 140 characters for a tweet, so if you exceed that (natively, not using another tool which allows more), you need multiple lines. The #sqlhelp hash tag is a great example where you may need multiple tweets to get an answer out.

This really begs the question: what is the best format for what you want to say? A single tweet? A post on a message board looking for responses that are more detailed? A blog post which will be more informative and detailed but not a tome? A whitepaper which can have more content than a blog post? Or do you just write a whole damn book?

Since I’ve done all of the above (in spades!), here’s some advice:

If you’re going to use the #sqlhelp hash tag with a complex question (for example, don’t ask for advice on a high availability architecture and expect a 140 character response that is definitive), put a link to a message board or somewhere similar (even a blog post) that people can respond in more detail.

As a responder – and I’m just as guilty on Twitter sometimes – having multiple answers is sometimes necessary, but if it gets too much, you may want to have the person contact you directly or blog about it. For me, Twitter is hard to follow versus a message board because of the lack of built-in, easy-to-use threading. As I know people do like to see resolutions on #sqlhelp because it may help them, it’s tough.

Having said that, I find #sqlhelp almost more immediate than, say, the MSDN or TechNet forums. There’s something very cool about that.

A tweet shouldn’t necessarily lead to a blog post or something bigger, but it may spark ideas.

If you like to blog (and who isn’t blogging these days?), the next logical step is something like a whitepaper. I find many people have short attention spans, so I try to keep my blog posts reasonable short (relatively speaking of course) and focused. Some people may do a series of related blog posts to break a topic up. That series may equate to a whitepaper. Personally, when I know I want to do a bigger topic.

Writing a whitepaper is easier than you think. Some tips:

1. Get it tech reviewed. Even the best of us have some things which are wrong or makes sense to us the way we word it but not to others.
2. If possible, get it edited by someone who does it for a living. It really does help.

That said, like a tweet or a blog post, sometimes you have a lot to say. I find most people’s tolerance for a whitepaper under 40 or 50 pages of text. If you’re writing a 181 page whitepaper, that’s pretty much a book.

Clearly a book is a big commitment. These days you can self publish if you really want. I’ve talked about it, Brent Ozar (blog | Twitter) has blogged about it, but don’t write a book if you expect to get rich. Do it because you want to and are passionate about the subject you are presenting. You employ all the techniques you would at the smaller levels, just on a much grander scale.

The problem I find with more instant media such as Twitter and blogging is that while it’s often right and good, there’s just as much out there that is bad and incorrect and can cause harm. Even if what is said is technically correct, it could be the wrong advice for some. You need to always put things in proper context. Before I say anything else, let me say I am far from perfect. I do get things wrong at times and if I do, I post corrections or note it in a way that people can see what the right answer is. The last thing I need is someone saying, “Well, Allan said this and it caused my <insert thing here> to fail.” Ugh.

The other mediums such as whitepapers and books tend to have reviewers who will ensure that things are right and the messaging is on point. Even that is a tricky process because if you get multiple opinions about something, it’s up to the author to make “the call”.

It’s great that we have all of this content out there and the ability to do our own – we just need to choose the right one for our messages!

February 10, 2011

Interview in Hebrew for the Israeli SQL Server Community (With More English)

I was recently asked to do an interview for the Israeli SQL Server community who has an excellent website.

The interview can be found in Hebrew at this link: http://www.sqlserver.co.il/?p=2579. They wrote a very flattering introduction.

Now, truth be told, I didn’t write it in Hebrew. If I was still in college, I probably could have but it’s been nearly 20 years since I’ve used my conversational and writing skills when it comes to the language. It would have taken a lot of time with the help of the Hebrew-English dictionary sitting in the other room. So to make it easy, they let me do things in English and translated it for me. Reading the re-translated answers in both Google and Bing’s translators are pretty funny to see some of the differences. Bing is worse. Try it for fun.

One of these days I hope to get over to Israel to do some consulting or teaching. I almost did a few years back but it fell through.

Anyway, here are the questions and my original English responses:

Hello, Allan. Please tell me about yourself and what you do.
Shalom, everyone. I grew up in New Jersey just outside of Philadelphia and now live near of Boston in the USA. I’ve been playing bass for over 30 years and music is still a big part of my life when I’m not working. I’m currently in the middle of recording a new jazz album that should be done by the summer of 2011 where I’m doing all of the writing and arranging. Although I don’t do it often, I can lead shacharit, mincha, and ma’ariv if I need to, as well as read Torah and Haftorah. However, my conversational Hebrew is not very good, so I’m having you translate this interview from English into Hebrew for me.

I’ve authored or contributed to quite a few whitepapers and books over the years. Most of the whitepapers can be found on Microsoft’s website. My latest book is Pro SQL Server 2008 Failover Clustering (Apress), and my previous two books were high availability books based on SQL Server 2000 (Microsoft Press) and SQL Server 2005 (Apress). I will be writing a SQL Server Denali high availability book as well. My blog can be found at http://sqlha.com/blog and my handle on Twitter is @SQLHA.

A few words about your background and career history.
I started with databases as a quality assurance (QA) intern working at Sybase while I was attending Brandeis University. That internship had three things that stuck with me: my interest in databases, SQL Server, and a passion for quality and the need for testing at all levels. From there, I held a few different positions, including a formal QA job, the database resource for a company where I had to deal with Oracle, Sybase, and SQL Server on multiple platforms, and finally a job as a consultant for Microsoft. From there, as they say, the rest is history.

Microsoft is where the doors opened to what I do today, including all of the speaking, training, and writing, and got heavily into SQL Server high availability. The first technical document I wrote for Microsoft that was published was the SQL Server 2000 failover clustering whitepaper. After Microsoft, I spent awhile at Avanade (the joint consulting venture of Accenture and Microsoft), and finally went independent 3.5 years ago and formed my company Megahirtz LLC (http://www.sqlha.com). I provide consulting, mentoring, training, content development, and other services for clients all over the world. I’m also honored to have been awarded a Cluster MVP from Microsoft. This is my second year as one.

What do you think about cloud computing for SQL Server in the upcoming years? Once SQLAzure is up and running (properly), would you join a facebook group we established, called “lay off all DBA’s and send them to Sibiria”?
I think there will always be a place for traditional database systems. In my opinion, right now the cloud space as it relates to SQL Server is still immature and has a way to go before it is mainstream. It will advance over the next few years. I don’t think companies who have certain constraints such as high security are going to be willing to place their sensitive data somewhere out there in the cloud where they have little to no control over it and who may have access to it since it’s on a system they know nothing about.

The cloud is similar to virtualization. Virtualization a few years ago was niche and a big topic of discussion as to how to deal with SQL Server. Today, nearly every company has some form of virtualization for their SQL Server deployments somewhere, even if it’s just the development or testing environments. It has its own set of challenges, such as determining performance bottlenecks when everything is virtualized, and I see the same going forward for cloud-based deployments. The cloud will definitely play a part in the way everyone in computing works going forward, but its role in how applications, databases, and DBAs act is still sorting itself out.

What it does mean is that DBAs and administrators of all kinds need to get smarter about what they do and keep their eyes open. You have to learn to adapt; you can’t just assume that the party is going last forever. This is why keeping one’s skills up to date and looking into the future is so important. I encounter customers that are 2, 3, or 4 versions behind in some cases and it will hurt them at some point, and it definitely makes the DBAs much less marketable to prospective future employers. When it comes to the cloud, DBAs cannot ignore it. If you have the right skills, you can adapt.

Which feature is the least useful in SQL Server history?
There are many features that have come and gone over the years that would probably qualify as not very useful. English Query is a good choice – did anyone even use that?

What does your typical working day look like?
It varies greatly. Sometimes I’ll be working onsite with a customer, other times I may be in a classroom training folks, or I could be home doing remote work or content development. I could have a typical 8-hour day, or one that is 15+ hours. One of the nice things about being a consultant is that the work is always different. I’m always looking to be challenged. I did write a blog post recently on the realities of consulting, and it has a bit of what I go through when I’m doing client work and on the road. That post can be found here: http://sqlha.com/blog/2011/01/26/the-not-so-glamorous-life-of-a-consultant/.

What is the last SQL Server book you read (not including yours…) ?
Honestly, I don’t read a lot of tech books in my spare time. I did buy Professional SQL Server 2008 Internals and Troubleshooting (Wrox), and I’ve read portions of it. It’s one of the better SQL books out there today.

Which feature in SQL Server or technology would you take with you to a desert island?
If I was on a desert island, the last thing I would probably be worried about is SQL Server. But since you asked it in a SQL Server context, I’d probably try to print out Books Online and take it with me so I could keep my mind active as well as learn a thing or two so when I got back to civilization, I’d still be useful.

What do you think of the future of SQL Server clustering and mirroring, given the new HADR technology in Denali?
Clustering is one of the parts of the new AlwaysOn feature. It provides the quorum/witness functionality. I’m not worried about finding a new job just yet! I’ve seen AlwaysOn for well over a year before it was publicly announced at PASS and it’ll be interesting to see what people think of it as they start to use it in the CTP releases. It’s definitely a bit more robust than mirroring was in 2005 and 2008. Traditional failover clustering is also fully supported in Denali and has been enhanced a bit when it comes to deploying a geographically dispersed failover cluster. I think database mirroring as it currently is implemented in SQL Server will most likely be deprecated, but it will still be around for a bit giving people time to transition.

What is the craziest SQL project you were involved in?
About 10 years ago, a large customer was having some severe downtime issues. A large team of consultants and support people were there 24×7 for months. Things did stabilize over time, but I can remember working a 26 hour day during that engagement. I also learned a lot, and it had a pretty big influence on starting me down the path of high availability and how to prevent downtime. There was never a dull moment on that project

And finally, a personal tip for the DBA in Israel?
Two things, one related to high availability and one not:

Achieving high availability is more than just a feature or technology such as failover clustering. You have to have processes, agreements such as recovery objectives, as well as needing the right people in the right place at the right time. If you don’t have any of that, technology won’t solve your availability problems. Technology should support the other things, not the other way around.
One thing I learned a long time ago: the only one looking out for you and your career is you. Whether your goal is to remain a technically-focused IT professional, someday manage others, or even do something completely different, take the right steps to learn and further your career so you meet those goals. Sometimes that may mean doing things on your own or even taking a leap of faith and believing in yourself to make a radical change. You may even need to take a step back to move forward.

L’hitraot!

February 9, 2011

Deck and Recording for the PASS DBA Virtual Chapter Webcast from Today

Thanks to everyone who took the time out of their day to attend my webcast “Windows Failover Clustering Basics for the DBA” for the PASS Database AdministrationVirtual Chapter. The recording should be posted up there soon under their archives. I had a great time and it was a lot of material to cover. Wish I had more time. I also apologize for the occasional audio glitch – not sure what happened. Towards the end we were getting some feedback as I was answering questions, so if for some reason it came out poor, feel free to contact me and ask your question again.

Here is the slide deck

I do have a multiple day failover clustering course which has labs, covers SQL, and I’m adding Denali content right now so if you’re interested, let me know!

February 9, 2011

Transitioning for the DBA Crowd

Whether you’re leaving for a new opportunity, a fellow DBA is leaving (or going elsewhere in the company), or someone you deal with that touches the databases you administer (such as the application admin) is moving on, you need to be prepared to do your own exit interview. Chances are their collective knowledge has not been sucked out of their head, and it walks out the door when they leave – never to come back. As a DBA, there are things you will need in this situation, or if you’re the one moving on, what you should provide people with so you’re not the guy they are cursing two weeks from now when you’re no longer there.

The Dos

If you don’t do it already, start documenting your procedures and specific things you do which no one else may know. For example, when the SuXor application had performance issues and you had to apply a hotfix to correct the problem, what was it? That info will come in handy if they ever need to rebuild the server.
Ask questions. What you don’t know will hurt you. Chances are you didn’t set up the servers since you just got into your DBA role 30 days ago, but you damn well better know what you have and what they need. At the end of the day, it’s your butt on the line when the CxO can’t get their report and it’s due to a missing linked server you didn’t know about.
Get passwords … and change them. If the person leaving set up the apps and databases, chances are they used a password which may not be documented somewhere. Get it. Even though whoever is moving on may have been the nicest of people, if you have a single password used for ‘sa’ on all of your instances, it’s best to change it when the person leaves. Even if someone won’t have VPN access, you don’t want to take the chance.
Return everything. Sure, the company probably won’t miss that $50 USB stick they bought for you and has no asset tag. Give it back.
Be proactive.If you know you’re leaving, initiate the transition of your duties, information, etc. In a best case scenario you do such a good job at it you can leave earlier than planned to start your new opportunity. People will hopefully remember you fondly if you do this.
Set a reasonable time limit for people to contact you after you leave or ask if you can get that person’s contact info. Look, you don’t still want to get calls six months down the road, but assuming you left on good terms, there’s no harm in answering some questions quickly over the phone or e-mail for the first few weeks after you leave. Good karma, baby. Set a line in the sand (or ask where that is). At some point if they still need your help, you’ll have to cut the tap off or tell them they need to start paying you. Just as much as it’s good to have a nice transition out, you still need to onboard at a new position. Understand if it’s not you and you get the “no more” message. Goodwill only goes so far.
Realize that life goes on without you. Hard to believe, right? Every company I’ve left has managed to survive without me. They couldn’t always replace me, but they managed to get on with it. You’re not the be-all-end-all.

The Don’ts

Don’t burn a path on the way out the door with a flamethrower. Whether it’s you or someone else, people on the other end don’t appreciate the person who is leaving doing things that make enemies. The person leaving may be unhappy, but there’s a pretty good chance he or she will run into those people again in the future. Bad word of mouth gets around.
Don’t destroy anything which may be of use to others. I’m sure your share on the corporate network has lots of good stuff in it. Don’t delete it without thinking. If you’ve got things people may need, send them in e-mail or ask people where to copy that stuff to. Believe me, they’ll appreciate it. Don’t assume they have what you have.
Don’t poach other employees. Chances are you’ve got a clause in your employee agreement which prevents it, but even if you don’t, it’s pretty bad form to do it. At least if you’re going to do it, let the body go cold and have some discretion so it appears they left of their own free will. The last thing you want is a lawsuit. I bet the corporate lawyers have time and money on their side.

So that’s what I’ve got so far … what are your Dos and Don’ts?

Popular Posts

Categories