I’m honored to be speaking at the next SQL Saturday in Christchurch on June 8th. Check out the link below:
https://www.sqlsaturday.com/831/eventhome.aspx
After taking most of last year off from speaking at these events, I’m looking forward to getting back into it, not to mention catching up with a lot of people I haven’t seen for a while. If you are in Christchurch that weekend be sure to come along.
I wanted to make sure everyone was aware of this great Initiative to support the Canterbury Muslim Community Trust. After the events of March I think a lot of people are left wondering what they can do to show their support of those effected and to help out. Steve Knutson and Hamish Watson have done a great job of pulling this event together.
Unfortunately I am not going to be able to make it down to Christchurch for this weekend, but I do hope that the extra link helps:
2019 has gotten off to a very busy start on the work front, but on the personal front it has been more of the same. I’ve just been putting one foot infront of the other, and I’ve been doing it alot. As a result I’ve dropped 10% of my body weight in the last 3 months.
Back in October 2018 I had a couple of health issues. The most obvious was a back injury which cost me a number of days off work and is still troubling me today when I move in the wrong way. Less obvious was a brief chest pain I experienced which I found a bit scary. My mother has suffered multiple heart attacks and my Grandfather died from one. While this issue was not a heart attack, it was scary. I knew I was overweight and working in IT I live a lifestyle which is mostly sedentary.
But while I wasn’t moving much, things were certainly very busy at work. Quite simply I felt that I didn’t have time to fit exercise into my life. I kept thinking “You can start getting fit when things aren’t so crazy” or “You can work on your health when this project is out of the way”. The only problem is, things are always crazy, and finishing one project just means you’ll have another one take it’s place.
So I took some advice from the Arnold Schwarzenegger when he was asked how he fit everything in. “Sleep faster”. I got a fitbit and resolved to getting up at 5 o’clock each morning and walking for a couple of hours. I still can’t run because it jars my back, but if I can walk, I walk. After a while walking every morning a few of those fitbit challenges seemed within reach and I found myself walking in the evenings at night. It’s a great way to unwind, get exercise and fresh air, and as a nice side effect I have slept very soundly since I started doing this.
Looking back at my history the first day I logged steps was November 21st when I was 98.8kg. Today I dropped under 90kg for the first time, weighing in at 89.6kg. That means dropping a kg roughly every 9 days, which was a bit slower than the 1 per week I was targeting, but still something I’m pretty happy about.
Why is this on a blog which is mostly technical and work related? Quite simply I’ve come to realize that everything starts with looking after yourself. I encourage everyone else to challenge their belief about how much free time they have to fit exercise into your schedule. My perspective has been flipped on its head and now everything else needs to fit around my walking, not the other way around.
I’ve been doing some testing on Azure database over the past few weeks. Obviously when you move to databases being used as a service you need to pay to get this service, but I did find a nice little workaround that unlocks the “free tier” for SQL Database.
The process is outlined below, but the TLDR version is – create a “WebApp + SQL” in the free tier and the database becomes free. There are some downsides to this of course – the free tier is limited to 5DTU and 32MB size, and it vanishes after a year. But for playing around those values should be fine – and if you need more than that it gives you a good marker for how much Azure Database will cost you, because don’t forget what you are paying based on in Azure DB.
To get your free Azure Database Click on “Create a Resource” and search for “WebApp + SQL”. Select it then click create and you’ll see something like this:
The free tier can be selected from within the Dev\Test toolset:
Then when you select the information for your database the super secret free Azure Database Tier has been unlocked.
Please remember the limits on that free database:
It expires after 365 days
It has only 5DTU
It cannot be larger than 32MB.
Microsoft have made the claim that you don’t need to run consistency checks on Azure SQL Databases because they take care of it. That’s nice, one less thing to worry about right? Of course to a DBA a claim like that falls into the same category as your backups being reliable without verifying them, or…I dunno…internet privacy not being a problem for people with nothing to hide. In short, unless I know HOW they are achieving this I am not going to consider my data safe, and even then I’m probably going to be dubious.
So it’s refreshing to see the blog post yesterday from Peter Carlin which outlines the lengths that Microsoft is going to in order to make sure Azure Database users are not affected by data corruption issues. Check it out here: https://azure.microsoft.com/en-us/blog/data-integrity-in-azure-sql-database/
The bit I particularly like is listed under evolving methodologies. It says “The Azure SQL Database engineering team regularly reviews and continues to enhance the data integrity issue detection capabilities of the software and hardware used in the service.” which is a nice acknowledgement that even with all the proactive work they are doing in this space, data corruption is a case of something going wrong, and there is always room for improving the method of dealing with it.
Thanks Microsoft.
Last Friday I had the very great pleasure of attending the Google IO extended presentation organised by Dhruv Valera. Dhruv put on a great event and there was a lot of interesting conversation forthcoming afterwards (along with a slight hangover). I found the content of what Google is doing absolutely fascinating, especially the concept of moving from a mobile first to an AI first approach. But the one thing that was missing was a coherent idea of what the future will look like. Google is making staggering progress, but it is incremental progress on where we are now, and I like to think the other way. Where do we want to end up, and how do we get there.
My own role is currently increasing the level of automation within our service offerings. When I get to do that it’s alot of fun. In one recent POC I automated 40 hours of work down to ten minutes, and even that 10 minutes was running time for automated processes, so the human interaction was around 30 seconds. That’s nice for a business, because they can charge 40 hours and pay for 30 seconds. And sure, there’s a bunch of development work that goes into that, and stuff can always go wrong with automation and all the rest of it, but one thing that is a bit scary is that 40 hours is a significant number. It’s a work week. Which means I can think about things in two ways. I can kick off this process and not bother showing up to work this week, log my hours, work bills the client and everyone is happy. Or alternatively, my employer can get someone else to kick off this process 20 times, pay them for 10 minutes work and there’s me and 19 of my co-workers out of a job.
When we talk about automation we don’t tend to talk about job loss, and in my particular business I don’t think that’s an immediate concern. I’m trying to fit everyones work tasks back into the 40 hour work week we actually get paid for rather than squeeze people out of jobs. There’s plenty of work to go around. But taken to it’s natural conclusion where do we end up? Back when I ran the internet cafe I used to joke that I could probably write a pretty simple program to do most of what I did. But what if that’s not a joke? Family aside, it’s quite possible for me to go through an entire week without interacting with another human being. I can pay at the pump with a credit card so I don’t need to go and see the service station employee, I can go through an automated checkout at the supermarket, I can order whatever I want online and have it just turn up. Where Internet Cafe’s do exist they probably have been fully automated by now. How difficult would it be to fully automate a cafe, or a bar, or a restaurant? Not too difficult. And even if the machinery to do that cost $50,000, or $150,000 – that doesn’t actually break down to that many annual wages. Just think if you owned a cafe and could spend $150,000 to never have to pay wages again… you wouldn’t need to think too hard. With minimum wage for a full time salary in NZ being a shade over $30K – That’s 5 employee/years and the equipment pays for itself. And you don’t have to worry about staff being sick, taking holidays, changing jobs, needing training, wanting raises and so on. It’s actually a pretty easy equation . Then you can add in the advantages – you could actually be open 24×7, because your staff never need to go home and sleep. You’d never need to recruit or train new employees, and you could scale your business up like crazy!
So when I see Google’s natural language processing advances, and see a phone that you can hold up to a plant and have it identified a bunch of stuff happens. The geek in me says “That’s so cool!”, but the philosopher in me starts to ask where it stops and how the world looks even 10 years in the future. For anyone who ponders the same questions I recommend you read Marshall Brain’s “Manna” – it’s free on his website and details two possible futures for mankind. The first is depressingly likely, the second euphorically enticing. Fair warning, if your brain works like mine you will also want to devour everything else on his website. It’s all fantastic stuff.
I am, often in the face of overwhelming evidence to the contrary, an optimist. I believe that fundamentally humans are good. I believe that technology should make us happier, smarter and more free. We can do amazing things if we set our minds to them, and the technology we have is going to make that easier and easier. So thank you Dhruv for organising the Google IO extended event in Nelson, and thank you Google for continuing to push the limits of what we are doing and can do.
Because I work in IT I quite regularly get asked to do any number of things that I have never done before on the basis that it involves computers(or increasingly phones) or the internet, or something….geeky. The idea is that because we work in one area of IT, our skills should neatly transfer to other areas.
This is an annoyance in your personal life. Of course I have no experience in how to set up photo sync on your iphone, I have an android, but actually, I can probably figure out how it’s done faster than the average user. Hook up your new playstation and transfer your accounts? I mean…sure…I don’t have an xbox, but I can read things on a screen and guess which button to press. But that’s quite low stakes stuff. The workplace equivalent is when a sysadmin or a web developer is expected to check in on the databases. In the former case they are probably looking at the database server as a bunch of files on disk, in the latter they are thinking of it as a bunch of tables to access to server data. They have become an accidental DBA and the expectation that has been placed on them is, quite simply, just not fair.
Databases are incredibly complex beasties, and they shouldn’t be treated as a bunch of tables, or a bunch of files. They are databases, residing on database servers and the skillset required to properly administer them is very specific and quite hard to define. But in this case I’m going to carve off the first 3 things to master if you find yourself lumped with the DBA tasks with one of those lovely little throwaway lines like “Hey, seeing as you have sysadmin, can you just make sure the sql servers are okay.” If you accidentally don’t say no to that question it will be the first of many questions that start with “Can you just….” Because everything about databases SEEMS like it should be quite simple to someone who doesn’t know how they work. Worse still it may seem quite simple to you too meaning you can do incredible damage very quickly if you are not fully aware of the implications of the actions you’ll be taking.
Number One: Backups
Okay… so when you said “I guess” to that you just made yourself responsible for the availability of data within your organisation. And you have to do that around your real job? Here’s the first thing that can go wrong – someone just accidentally overwrote the database that controls all the company orders – you are about to get one of those “Can you just?” questions. “Can you just restore that back to this morning at 10:36?”
Well can you?
The first skill you need to master is setting up SQL backups. That’s so much more than just running a backup command. You are not just creating a rollback point prior to a deployment, you now need to understand what the business expectations are around recovering data and how to implement them. I’ve already dedicated a bunch of posts on this blog to how you go about doing this, but I’ve barely scratched the surface. The key bits:
Which of those is the most important? The next one on the list. You need to have done ALL of these before you should feel comfortable with your ability to restore data in your environment.
Number Two: Monitoring
What do you do if one of those backups you just spent 3 months getting everything in place for fails? Do you even know it happened?
SQL Server has a wonderful selection of built in tools to help you out. You want to make sure if something goes wrong you know about it. Here’s a selection of SQL components you may want to look into to make that happen.
Now rolling your own monitoring can be a time consuming process and somewhat prone to error. My advise is never set up alerts that come to yourself. Set them up to go to a generic address like DBAalerts@yourcompany.com. It’s fine to have that simply forward everything on to you, but you want to be able to let whoever takes over from you to simply be able to start picking up alerts, not have to go through and manually change everything you have set up.
Number Three: Automation
I’ve already talked about a number of processes and tools that you can set up once and then only need to revisit if something goes wrong. The old fashioned DBA would log into each server and work through a checklist of tasks, and that just sucks. One of the things you will find very quickly as an Accidental DBA is once people know you are the one who can make things happen on the SQL server you will be hit from all sides with all sorts of requests. Your time will be too scarce and valuable to waste on something you could have scripted once and set to run on an appropriate schedule.
Learn the power of SQL Agent for scheduling tasks. Add powershell and command line tools to take care of the associated operating system tasks and make your life easier by overseeing an army of robots set up to complete your tasks and make you look good.
Number Four: Meet the Community
SQL Server has, without doubt, the most motivated and engaged tech community I have every been involved with. Go and meet them. Most reasonably sized cities have a SQL user group. Microsoft has regular events in large cities. But even if you can’t engage in these in person, the online community is vibrant and full of people who are willing and able to help.
Sometimes that help is in the form of answering questions(I think you are more likely to get answers quickly to SQL Server questions than most any other technology I have been engaged with), sometimes it is detailed blogposts\instructions on how to perform more complex tasks, and sometimes it is the perfect community tool for the task you are working on. If you haven’t already go and check out sp_whoisactive, Ola’s awesome maintenance scripts or DBATools. Every one of these is commercial grade brilliant and they are all available, along with a community that will tell you how to use them….for free!