Today is the day that I have the privilege of announcing the PASS Summit 2018 Speaker Idol contestants. The contestants have been grouped into three groups, with each of groups presenting on either Wednesday, Thursday or Friday.
The groups that we have for the PASS Summit 2018 this year are:
Wednesday (4:45 PM)
- Heidi Hasting
- Ameena Lalani
- Christopher Wolff
- Dennes Torres de Oliveira
Thursday (4:45 PM)
- Leslie Andrews
- Janusz Rokicki
- Michael Johnson
- Paresh Motiwala
Friday (11:15 AM)
- Randolph West
- Peter Shore
- James Donahoe
- Rob Volk
The winners of these three days will then be presenting again in the finals on Friday afternoon at 3:30 pm along with one runner-up. Each of these 12 attendees to going to get some great feedback from our panel of judges (which I’ll announce later).
I wish everyone luck, and all see you at the PASS Summit.
If you work in IT, you probably heard about the disaster that Azure had last week with their South Central US (Texas) region. This outage has been a significant problem for a lot of customers, as well as to Microsoft.
When it comes to having a high availability or disaster recovery plan in place, they become vital very quickly. Based on what we know so far, this started with a lightning strike. A lightning strike wasn’t exactly something that could be planned for or could have detected in advance. Because we couldn’t prepare for the problem, all we could do is react. Reacting to an issue like this requires planning. That planning is a High Availity or Disaster Recovery plan.
On social media, I see questions like, shouldn’t Microsoft be doing High Availability and Disaster Recovery for you if you are in Azure? My answer to that is a resounding no.
High Availability would only be useful for a full building failure like Microsoft Azure had, if it was set up recently using Zones. The High Availability would only be successful because a single building was taken offline due to the lightning strike. The outage only impacted VMs, and not entirely. The outage leads me to assume that PaaS services were able to stay online because the PaaS services are set up to use Zones within Azure. If VMs aren’t set up in Zones, then the services that they provide have no guarantee that they would stay online.
If you’ve told customers that Azure VMs will be staying up, without building High Availability into the design, then you’ve told the customer something that simply isn’t true.
Disaster Recovery isn’t included in any service that Microsoft offers unless you pay for it, and even still Disaster Recovery planning is only available on a couple of services that Microsoft provides. As Microsoft doesn’t offer Disaster Recovery on most of the services, this means that you need to plan for your Disaster Recovery needs yourself.
Yes, Disaster Recovery planning costs money, I get that. These costs might be hard for a company to swallow, but how long can the company last if the services that it depends on are offline. Can your company last for a day or two with customers getting no response to your website? These customers might assume that you’re gone out of business and use a competitors service instead of yours. That might mean that money is lost, not just now but in the future.
Disaster Recovery doesn’t have to be a full disaster recovery environment. You may be able to get away with restoring backups to get the website back up and running. You might be able to get away with a small part of your environment being set up for Disaster Recovery. You might need a full Disaster Recovery environment setup. It all depends on what your specific business needs and what you can afford.
All Clouds Can Fail
If your applications aren’t in Azure, or Amazon’s AWS service, or Google Cloud you aren’t immune to this problem. If you have a server at the office that hosts the application that runs your business, what happens if a lightning strike happens to your building and the server you have stopped working? Will your business survive? What are the plans if this happens? If you don’t have this plans, or you need to review these plans, we can help, you need to reach out to us, and we’ll see what we can do to help you out, including what kind of High Availability and Disaster Recovery plan is right for you.
Yes, High Availability and Disaster Recovery is expensive, or it can be. But when the time comes, it’s costly not to have High Availability and Disaster Recovery.
The PASS Summit 2018 is quickly approaching. Every year the PASS Summit has lots of new attendees, and the PASS Summit is large and can be intimidating for a new attendee that’s never been to the PASS Summit before. Because of this, I’m presenting the annual PASS Summit Attendee Orientation.
This webcast will be full of great information that you’ll want to know in order to get around the PASS Summit and the city of Seattle in general. Attendance is a must for the new attendee while returning attendees will find the information that I’m presenting useful as well.
So get signed up, and I’ll see you on the 2nd.
Another excellent PASS Summit is almost here. And with the PASS Summit comes the SQL Karaoke party. And once again DCAC is hosting this great event which will let us all ring in the PASS Summit in style. We have two fabulous sponsors for the event this year, with our great friends at SentryOne and SIOS sponsoring the event.
Just like in years past, this is the PASS Summit party you don’t want to miss. We’ll have the live band playing all your favorite tunes for you to sing to, soda, beer, wine and well drinks for those with a wristband and a cash bar for those that don’t get a wristband.
The party starts at 9:30 pm and rocks until 1:30 am giving you plenty of time for a nap before the keynote begins on Wednesday.
Be sure to bring your Eventbrite ticket to the door to get your wristband, as well as Photo ID (drivers license or Passport) to get to get into the bar). And come and have a great time.
The PASS Summit Summit code of conduct will be in effect at this event.
You’ll find the link to the party here, so get registered.
Just like in prior years, you should register for EITHER the cash bar or the open bar. The open bar includes beer, wine, and well drinks. For the cash bar, you’ll be paying for your drinks. So get signed up, and we’ll see you at the party.
Recently on Twitter I saw a question about what index maintenance you should be doing on your ColumnStore Indexes when it comes to maintenance. As I looked through DCAC’s blog posts I realized that none of us had blogged about this topic before.
The reason for this is pretty basic, there’s really nothing to do when it comes to maintenance on a ColumnStore index.
Normally we do maintenance on indexes to reduce fragmentation, but there’s no fragmentation on ColumnStore indexes as the data isn’t stored on the table in a sort order. Data pages that aren’t needed anymore are removed automatically (the same with dictionary pages). The other big thing that index maintenance does, it update statistics on indexes. Well, there are no statistics on ColumnStore indexes, so there are no statistics to maintain. Why aren’t their statistics? Because every operation is a scan of a ColumnStore index. Since everything is a scan, there’s no need for statistics.
I admit it, I still use SQL Profiler. I always have, and I will for the foreseeable future. My reasons are pretty simple.
- When it comes to troubleshooting I can spin up a Profiler session must faster than an Extended Events session.
- Usually, I’m troubleshooting something as a one-off. So having my session isn’t really something I care about.
- I usually can’t bring scripts into my client’s servers to setup Extended Events, so I have to go create everything manually. (See #1)
- Done correctly with filtering, there’s no risk of a production outage using Profiler.
- If I could bring in scripts to set up an Extended Events session (see #3) I’d have to modify the session which I can do faster in profiler than in Extended Events.
Should you be using Extended Events? Probably. Odds are you got a full-time job somewhere, so extended events sessions are going to make more sense for you as you can run them against your servers and easily jump on and see what the server is doing.
What would it take get me to use Extended Event sessions instead of Profiler? Speed. Whatever GUI Microsoft creates for Extended Events needs to be just as responsive as the Profiler GUI, and the data that is returned needs to be returned by Extended Events just as quickly as data is returned from Profiler.
If you’ve been thinking about submitting for the PASS Summit 2018 Speaker Idol, but you aren’t sure that it’s for you, we can solve that problem. On
August 30th, 2018 at 1600 UTC (4 pm) join myself, Karen Lopez, and Joey D’Antoni at the PASS Professional Development Virtual Chapter as we talk about Speaker Idol at the PASS summit. You can register for the session on the PASS Professional Development VC website. We look forward to seeing everyone at the webcast.
While yes, the PASS Summit is a few months away still, the window for submitting for the PASS Summit 2018 Speaker Idol competition is starting to close. As I announced in June all entries to this years speaker idol need to be submitted by September 8th, 2018. That’s just a little over a month away. This give me time to review the submissions, notify the contestents, have a call with them, and give them time to make their presentation for the PASS Summit.
If you’re an aspiring presenter who’s looking to learn from some of the best presenters out there, then this is for you and you need to sign up.
Azure and the load balancers can be annoying little things. Especially when you are doing maintenance on your Azure environment, or setting it up for the first time. One of the quirks you may run into is that if you leave an Azure Load Balancer sitting for to long with no machines behind it, you may find that the load balancer itself just stops working.
Thankfully fixing this is actually pretty straight forward. Simply delete the load balancer from Azure and recreate it. Now you may be worried about dropping the IP address that it’s configured with, and don’t worry, you don’t have to. If it’s an Internal Load Balancer (for a failover cluster or a SQL Availability Group for example) then it’s just a private IP and you can just reassign the same IP when you recreate it. If it’s a public IP, you don’t need to drop the public IP address object from Azure, which means that the public IP address will stay right where it is. You can just reuse the same IP address object and you’re good to go.
It’s an easy fix to an annoying problem. For internal load banacers (ones with private IP addresses) they can be really hard to troubleshoot as there’s next to no logging done on an internal load balancer.
A really great feature in Azure SQL DB went GA today. That feature gives you and SQL DB the ability to automatically fail databases over to a Secondary replia, without having to configure your application to handle that failover. You point your application at a VIP and that VIP will automatically handle failover of the resource.
Say for example you have a database in US West named db1-west.database.windows.net and the DR copy of it in US East named db1-east.database.windows.net. This feature lets you create the VIP db1-vip.database.windows.net which automatically points to whichever database is currently active. In the event of a failover of US West, the VIP is going to failover to the database in US East, the database in US East become writable and when the US West is back up, the data will sync back.
Another cool thing which this feature does is something that most features won’t do, it’ll trigger a failover that allows for data loss. Now, this normally would be a very dangerous thing, but the Azure team has come up with a safe way of doing it. When you figure the service to do the failover, you decide how long you want to wait for there to be no data loss. If you want the system back up as soon as it allows for, select the smallest number, otherwise select a larger number. This allows you, and the business unit that you support, to decide what level of protection you want to have built into the system.
If you are thinking about moving to PaaS, not being able to have a DR option may have been stopping you. This is no longer a blocking point, you now have an easy to configure DR, that you can manually failover is need be. If you’re thinking of moving to Azure, DCAC can help to plan and execute that migration. Contact us today, to schedule a meeting to discuss if the cloud is right for you.