Episode 70: SQL Server on Linux

Episode 70: SQL Server on Linux

Episode 70: SQL Server on Linux 560 420 Carlos L Chacon

In the keynote of PASS summit 2016 Rohan Kumar, the program manager for SQL Server proudly displayed a database restore to a SQL Server instance installed on Linux.  What makes this all even more interesting is the database was sourced from a Windows server.  While it might not sound like much, the number of operating systems SQL Server runs on has now doubled.  This is no small feat and our guest today Travis Wright is the program manager for the SQL Server on Linux migration.  While I have to admit I was not super excited about the news when it came out last year, I am definitely more interested and think there is a huge opportunity for SQL Server administrators to get access to jobs they wouldn’t have before.  I know you will enjoy this interview with Travis.

Episode Quote

“Why are we doing this? . . .  It comes down to the customers have been asking us to do this for a long time.” Travis Wright

Listen to Learn

  • Why are they doing this to us? (come on, that is funny)
  • What features are going to be supported.
  • The version of Linux supported.
  • What type of integrations are now possible – think WordPress on SQL Server.
  • How do I kick the tires on this?

Travis on Twitter
SQL Server on Linux
Data Platform insider blog
SQL Server on Linux Preview

About Travis Wright

SQL Server on LinuxTravis Wright is the program manager for the SQL Server transition to SQL Server on Linux.  Travis is the father of 3 kids and is currently on his second tour with Microsoft.  He was heavily involved in the system center team and is now putting on his skills to us on the Linux team.  He describes himself this way–“In a room full of “suits” and “propeller heads”, you can find me in the middle fluently talking about business strategies, priorities, requirements, financing, and schedules as easily as software architecture, algorithms, and database schemas. My specialty is bringing together the business flint and software steel to ignite the spark of innovation that creates something meaningful.”

*Untranscribed introductory portion*

Carlos L Chacon:               Travis, welcome to the program.

Travis Wright:                    Thank you. Glad to be here.

Carlos L Chacon:               Yes. Thanks for coming with us on this show … Thanks for coming on the show with us today. One of the very interesting things that happened in the key note at the SQL summit in Seattle was a demonstration of a restorer of a database that was a window sourced database and was restored to a SQL server running on a Linux machine. Lots of chatter about that. Ultimately, we wanted to have you on the program to talk a little bit about why this move to SQL server on Linux? I must admit I’m not a huge Linux person myself, what Microsoft sees as the future, I think there’s going to be some analytics discussion in here. Let’s go ahead and jump in and tell us a little bit about that history. Why the change forth to SQL server on Linux?

Travis Wright:                    Yeah, this is one of the first questions we always get is why is Microsoft doing this. After all these years you mentioned that you’ve been working on SQL for 26 years, why after all these years is Microsoft doing this. Really it comes down to the customers have been asking us to do this for a long time. For various reasons, it’s always been either a technical challenge that was seemed insurmountable or there was political challenges to do that within Microsoft obviously with Windows being a huge part of Microsoft’s culture.

Carlos L Chacon:               Sure.

Travis Wright:                    Everything. The star aligned recently though about a year and a half ago. I’d say we revisited the decision to do this, went on did a bunch of research talking to customers about what they’d like to see, talked to some developers. What we found is very interesting that Linux is very much a growing and vibrant ecosystem. We’re getting up to the point now where a SQL, Linux VM and azure accounts for about 25% of the total BMs and azure even. Lots of interesting things are happening on Linux. We got lots of big data things happening, a lot of CICD types of things are happening on Linux. It’s just a very interesting ecosystem. It’s growing. Even if you look at the IDC numbers for the database market you can see the Linux is growing faster than Windows as far as it being a database platform. Over time those markets will be roughly equivalent in size by about 2020-2021, somewhere in there. Really it just is a matter of being responsive to our customer’s requests. Looking at the market opportunity ahead of us as well as for our partners, just the ecosystem in general. Just doing what customers want. What customers want is to have a choice. SQL Server is really become a very mature database at this point. It’s full featured. Recently we’ve surpassed Oracle on the Gartner MQ For the first time.

Carlos L Chacon:               Hooray.

Travis Wright:                    We have retained that crown if you will for another year here recently. Now it’s really just a matter of; okay, should we go and add yet another feature of the SQL server which would of course be great or should we maybe take a look at expanding the availability of SQL Server to other customers, developers, and just industries in areas that we haven’t been in before. Just take the goodness of SQL Servers to those other areas. There’s lots of other reasons too. We’ve talked to customers, sometimes for whatever reason, customers want to run their database on Linux. The reasons range from that’s what their database and systems administrators are experienced in. They have Linux people, they want to use SQL Server. “Okay let’s do it.” In other cases you have companies that have maybe been acquired by another company that happens to run all their database platform on Linux. That acquiring company is now requiring the acquired company to map to that.

Carlos L Chacon:               Sure.

Travis Wright:                    The acquired company want to keep their SQL Servers. These types of scenarios come up where customers really just need this flexibility to be able to run SQL Server on other platforms besides just Windows.

Carlos L Chacon:               Am I naive in saying that that’s more than just hosting companies because I could see the hosting companies doing that. Obviously as they consolidate and things like that. Are there … You’re still seeing other places, just takes some general verticals like even healthcare or engineering or things like that are also wanting to do those the same changes?

Travis Wright:                    Yeah, absolutely. Lots of enterprises that are running their structure, whether it’s on or private cloud, or a host of product or something along those lines where they’re in their own SQL Server instances are very interested in this. It’s interesting, I took a trip down to Silicon Valley and visited with several customers down there where Linux is very hyped. They were interested in how they could use SQL server in those situations as well. Definitely lots of enterprises. Then as you mentioned hosters are also very interested in this because typically hosting providers are based on Linux. We’ve had a number of hosting providers approach us about now operating SQL servers as a database platform in their portfolio because we have now support for SQL server running on Linux.

Carlos L Chacon:               Exactly.

Steve Stedman:                A question on that then, because I’ve heard rumors of Ubuntu support, or Red Hat support, is there anything that you can talk about with what Linux distributions that you will be supporting?

Travis Wright:                    We’re starting out with support for Red Hat Enterprise Linux. The version that we’re supporting there is 7.2. We also support Ubuntu 1604. We’re now working on support for you SUSE Linux Enterprise server version 12SPQ. Those will be the 3 main Linux distributions that will support the SQL Server on Linux. We’ve had requests from customers to support some other distributions like Sento-S or Oracle Enterprise Linux, Amazon Linux, so on. I think as we wrap up the work to support those first 3, we’ll start to look at other distributions that we may want to support and working out what that might look like.

Steve Stedman:                Okay, when you talk about the Red Hat Enterprise Linux for instance, to be running on, is that something that a developer or someone who’s just trying to a proof of concept to try out on their own machine using just regular Red hat or does it obviously have to be enterprise there for people to use it?

Travis Wright:                    Right, Red Hat has a developer program. People can go and get Red Hat licenses that are intended for use in a development environment. It doesn’t grant you the right to run things in production. You can still get all the documentation and the subscription access to be able to get the packages and that kind of thing as a developer.

Carlos L Chacon:               Okay. Great. I have to check that out. So then when we talk about the Linux edition I mean I’ve heard of this being like core engine features. I mean will this just be the SQL engine or will it have other things like SSAS or SSRS or integration services or anything like that?

Travis Wright:                    Right. The first release we have of SQL Server on Linux will be focused on the relational database engine. That includes all the features around data being stored in a relational quantity. It’s all the DDL, schema, then storing the data, that kind of thing. It also includes a lot of the advanced features that we’ve been introducing in the more recent version the SQL server. Like in memory LLTP, column store. Lots of the other goodness of SQL Server; like compression, partitioning, and all those things. It includes all of the new security features we’ve been adding as well. Things like always encrypted to track and granted encryption, level of security, date of masking, auditing, all of those features as well are also included. Also things like SQL Server Agent, we can get in the high availability here as a specific topic here in minutes. All those features around the core relational engine. What’s not included in the first release is things like Integration Services, analysis services, reporting services. Those are the big areas that we’re not going after yet. Then some other miscellaneous items or things like master data services, data quality services, some of those kinds of things. Over time, we’ll be responsive to customer feedback. We’ll a look at what it is that people want and go after those things that make sense. I think integration service is probably the one that pops out the most as something the customers are looking to be able to run on Linux and so we’ve started looking into that and working on that. We’ll see what the time frame will look like.

Steve Stedman:                Okay. It really sounds like it’s the core engine but it’s going all the way with the core engine.

Travis Wright:                    Right.

Steve Stedman:                As far as the core SQL functionality there.

Travis Wright:                    Exactly, you should expect nearly all of the features that you would see in the co-relational engine to be there.

Carlos L Chacon:               Yes, when you think about that SQL server service.

Steve Stedman:                Okay.

Carlos L Chacon:               Everything would be there. Lets go back, you mentioned you broke out the high availability solutions, should we dive into that for a second?

Travis Wright:                    Yeah let’s talk about that. It was different levels of high viability and disaster recovery. Lots of different features around that as well. We just finished up the work on the SQL server agents to enable a lot of shifting. That’s one option. You can log shift from a SQL Server running on Linux to another SQL Server running on Linux or a working SQL Server around Linux, you’re supposed to run on Windows, you kind of have some flexibility to go about how you log shipping. That’s one option. I’m just going through the basics to the more advanced.

Carlos L Chacon:               Sure.

Travis Wright:                    You also have support for backup and restore. You mentioned at the top of the session here the demo that Ro-On did at the past Summit where he showed taking a database that had been backed up on Windows and restoring it on SQL Server on Linux. At the other direction as well. You can go take a database back up on SQL Server on Linux and put it on Windows. You have old backup, restore, attach, detach capability because of the binary compatibility between the SQL servers regardless of where they’re running at. That gives you a complex ability there. Then we also demod at one of the sessions we did at past. The fail over clustering running on Linux. We had 2 different notes in a cluster using pacemaker which is an open source clustering technology. I had to coordinate that cluster. Then we had to shared disc running on another VM. We showed how the SQL servers were using that shared storage to have all their data files, the master data as well as user data bases all right there. Then we would use the fail over command line utility for pacemaker to fail over from one node to the other. I had a little application that was running there. It was written in node. It would just sit there and run a simple query against the database nonstop. I would kick off up that application, it would sit there and would output the name of the server that it was connected to. I just output that to the terminal just every second or so. I failed over and you could see how it’d lose the connection for a few seconds. It would pick up that connection again and output the name of the new server to the terminal. It was a cool demonstration of how fail over clustering just works. Very similar to how it does on industrial cluster services, in this case running on Linux using pacemaker.

Carlos L Chacon:               Sure. That’s very cool because, again, I think sometimes maybe we forget until we get into the thick of it. Particularly in availability groups even more so. That reliance on the Windows server cluster service, right? Through to manage some of that. It’s very cool that there are alternatives in the Linux world to do that.

Travis Wright:                    Yeah, absolutely. Going on to the best of all worlds is where you start to get into availability groups and having high viability that way with the replication based HA. That, we are also working on. That will also be based on pacemaker. We will have that available for people to test here pretty soon as well. That will be pretty exciting because then you going to have the replication based HA using always on availability groups with reader election and fail over detection, being able to detect a node being unavailable and automatically failing over a lot of listener. All those things that you’re used to in always on availability groups on Windows, those will also exist on Linux as well, just using a different technology. One of the most interesting things about the way that we’re doing higher viability on Linux is that it will be a more open architecture. What I mean by that is that on Windows today Windows Server closer services is the main way of doing clustering in high availability with SQL server. On the Linux sides, there are lots of different technologies that are used for clustering in the Linux world. Pacemaker happens to be one of those and the most commonly used. There are others like zookeeper, there are some harbor B solutions. That we want to do is we want to make this a more open architecture where people can introduce other higher availability solutions that SQL server can work with. People can just write the scripts and talk over APIs, just introduce other solutions for high availability beyond what we provide as a reference in founding the recent pacemaker.

Carlos L Chacon:               Yeah. Wow.

Steve Stedman:                Interesting.

Carlos L Chacon:               You’re making it hard for your documentation folks there.

Travis Wright:                    Yeah, we just hired 10 more. It was starting to be a challenge to keep up with the documentation in general. You’ve got features that are going to work here and not there. You’re going to have what’s applicable to Linux, how you do things here versus there. There will be a ton of overlap. I’m expecting there to be just, if you had a chance to play with it, I don’t know yet or not. Most things just work, right?

Carlos L Chacon:               Sure.

Travis Wright:                    If you run an application, that just sits on top of TBS, it just works. All your drivers work. Anything that you’re doing through the front door, that all just works as you would expect it to. You got 2 SQL statements, most of those things just work. DMV’s just work. You can use SSNS, you can use SSPP, all those things just work. There will be some differences. Yeah, the documentation people have some good job security for the next couple years keeping up with all this. Yeah.

Steve Stedman:                Wow. Impressive for the initial release on Linux. I’ll tell you that. I know that for years Marilyn has been announced that it’s going to be deprecated. I would assume that Marilyn is probably not included in Linux. Is that correct?

Travis Wright:                    Right. Exactly. We’re not planning on bringing Marilyn over to SQL Server on Linux.

Steve Stedman:                Okay. I know you mentioned replication around availability groups. You probably weren’t talking about replication in a publisher or subscriber mode, were you

Travis Wright:                    I wasn’t when I was talking about earlier. Let’s hit on that briefly.

Steve Stedman:                Okay.

Travis Wright:                    With SQL Server agent working now, we do have sub-subscriptions as well working. Most of that just works. I think the only thing that is left to do in that area is where you’re doing some publishing and subscribing between different database engines; like SQL Server and Oracle for example. We need to figure out how we host third party binaries in our platform on Linux before we can get some of those working.

Carlos L Chacon:               Sure.

Steve Stedman:                Okay.

Travis Wright:                    As long as the SQL is looking good. We’ve been doing some testing with that. That’s doing well

Steve Stedman:                Great. Not that … I mean, it’s one of those that a lot of people are using it but the whole subscription piece is not … Publication piece not a piece that a lot of people get excited about. It’s being used in a lot of places.

Travis Wright:                    Yeah, it’s one of the things that are around for a long time. People depend on it. We’ve got to make sure that works.

Steve Stedman:                I know with the Windows version of SQL Server, there’s generally been the Express, a standard, and the enterprise edition. Although there’s been a few other little flavors over the years. Will the Linux version follow a similar licensing model or will it be something completely different.

Travis Wright:                    Yeah. The way we think about it is that SQL Server is the product. You license SQL server. Now you’ll just have an option of where you run SQL server. Could be on Windows Linux, could be on Docker containers. To that end, we plan on having the same edition lineup on Linux as we have on Windows. You’ll have a developer edition, you’ll have an enterprise edition, standard, express, Web Edition and so on. One thing that we don’t currently have a plan for right now is local DB on Linux.

Steve Stedman:                Okay.

Carlos L Chacon:               That’s your U-fold curve all of a sudden. You just said local DB.

Travis Wright:                    What?

Carlos L Chacon:               I guess I’m drawing a blank. What does that mean?

Travis Wright:                    Local DB is an optional component you can install as part of express. It runs as a process instead of a service. It’s just a lightweight database can be used in the context of developments. Typically few people use it like in Visual Studio. It’s a little lightweight DB that they use for development.

Carlos L Chacon:               Sure. Okay, I got you.

Travis Wright:                    Yeah.

Steve Stedman:                As far as the performance goes with this, one of the things that I’m wondering … I mean the initial version around the performance is always a big deal. Are there any metrics out there on how it will perform to comparing SQL Server running on Windows versus SQL server on Linux using the exact same hardware

Travis Wright:                    Yeah, great. We haven’t published numbers on that yet. We’ve been spending the last couple of months really focused on making sure that SQL server on Linux performs in scales and that kind of thing. Of course SQL server on Windows is the most obvious benchmark that we can measure ourselves against because that’s really apples to apples. We also do use some bench-marking relative to other database engines and like that. At this point I would say that we are at about 90% or so of the performance of SQL Server on Windows on a machine that’s a 2 socket machine with somewhere around like 256Ghz of RAM, something like that right. We have lots of work still to do in this area. We need to add support for Numa. We have some other improvements we want to do around network IO and some other code path improvements. There’s lots of work still to be done. Our goal is to get to the point where we have essentially the same performance and the scalability is what people are used to today with SQL Server on WIndows. We want to make sure that SQL Server on Linux can support those same mission critical to your one-type workloads that people use SQL Server for today.

Steve Stedman:                Okay, excellent.

Carlos L Chacon:               So is SQL Server going to be open sourced?

Travis Wright:                    Yeah, that’s maybe the next logical step, right? I think at this point, there’s no plans to do that. I think that would be a massive development effort on our part to get SQL Server ready to be open sourced even.

Carlos L Chacon:               Yeah.

Travis Wright:                    Let alone the will it would take to do something like. First steps first, we’ll get SQL Server on Linux out there, we’ll go from there.

Carlos L Chacon:               Very nice.

Steve Stedman:                Okay. One of the things today that I see is if somebody has a website they’re hosting with Word Press or something like that, the backend out there is generally a MySQL database.

Travis Wright:                    Yeah.

Steve Stedman:                Do you ever see SQL for Linux competing in that area with maybe an Express Edition or something?

Travis Wright:                    Yeah, absolutely. In fact, we’ve been working with some guys that started this project called Project Nami. Have you guys heard of this?

Steve Stedman:                I haven’t.

Carlos L Chacon:               I have not.

Travis Wright:                    Yeah. Project Nami, N-A-M-I, which funnily enough we’ve been talking to them for a couple months, then one day we’re finally like, “What is this Nami thing? What does this mean? They said it stands for ‘Not Another MySQL Instance’. What they’ve done is they’ve created a compact layer for MySQL … I’m sorry, for Word Press to run on SQL Server.

Carlos L Chacon:               Interesting.

Travis Wright:                    Very.

Steve Stedman:                Wow.

Travis Wright:                    Their original intention was to create this such that it could run on top of an azure SQL database which is where most of their customers today are using it. It also, because we have this compatibility between azure people database and SQL Server and windows and SQL Server on Linux, it happens to also work with SQL Server on Windows and SQL Server on Linux. They’ve taken it now a step further to where they have the full stack of Word Press and SQL Server running on Linux. A very interesting scenario. As we hit public review, we’re looking to how we can put that out there and make people aware of it. That certainly could be a very interesting scenario for people to run a Word Press instance on top of a SQL Server Express or standard. Depending upon what size they need to have their Word Press application or blog running on.

Carlos L Chacon:               Right now, I think that would be very attractive. I know, I was with a client that SQL Server shop, their marketing department wanted to move to Word Press. Which meant that all the sudden they inherited some MySQL databases that they now needed to care and feed, pure Microsoft’s shop if you will. That was just another item that they’re going to tackle.

Travis Wright:                    Sure.

Carlos L Chacon:               Being able to keep data all in the same family if you will, I think, would definitely be of benefit.

Travis Wright:                    Yeah, and definitely check out this Project Nami thing. It’s an open source project. I think it’s amazing actually. Very cool.

Steve Stedman:                Yeah. Definitely have to take a look at that. I know every time I try and use MySQL behind what I have in my blog, I just feel completely impaired trying to use it.

Travis Wright:                    Yeah. You want to know what you … You want to use what you know.

Steve Stedman:                Right, you know? Okay. Then for someone who wants to try it out; once you’re at that preview point, whenever that may be, how do you go about really someone getting started with that? Let’s say someone has a little bit of Linux experience. Are there going to be some ‘how to’ guides or anything like that that people could just walk through and say, “This is what it takes to get it going.”?

Travis Wright:                    Absolutely. It’s actually super simple to get going with SQL Server on Linux. I think we’ve done several demos now at conferences of the installation experience. It’s really quite refreshing for people that are used to installing SQL Server on Windows. On SQL Server on windows, you go through a wizard. It’s fairly simple and easy to do. It does take some time to click through everything and answer all these questions and everything. With SQL Server on Linux, we’re leveraging the package management systems in Linux. On Ubuntu, you have APTs, on Rel you have Yum, on Suso you have Zipper. They have all these packages. Installing SQL Server on Linux is really a simple command that you run at the terminal. Like on Rel, for example, you would say Yum install Ms SQL Server, it would go through the entire installation and download experience in about 40 seconds. That’s all it takes to get SQL Server installed. Then you run this setup script. It prompts you to accept the EULA and provide the SA password and you’re done. All in all, in less than a minute, typically you can have SQL Server downloaded, installed, and running. A very cool experience with that. Then I’ve recently just been falling in love with Docker. I don’t know if you guys have had much experience with Docker. With Docker, again, it’s a very simple acquisition experience. You just go to the terminal, assuming you already have Docker installed and configured, you just go to the terminal, you type in Docker Pull and their SQL Server. It will pull down the latest version of SQL Server, that image. That takes a minute or 2, depending on how fast your network connection is. Then you just type a command Docker run and their SQL Server, you pass a couple of parameters to it to specify the SA password and accept the EULA. You’ve got SQL Server running in a container on your mac book or on your Linux development environment, or wherever you might be running Docker. Then that also is just a beautiful thing when you start using some of these container management platforms like Red Hat Open Shift, For example. We just did a demo about a pass where people can publish that image up into their catalog and open shift. Then they can just provision containers just with basically a one-click experience in the open shift portal and push those containers out onto the container management platform.

Steve Stedman:                Very cool stuff.

Carlos L Chacon:               I can …

Steve Stedman:                I know … Go ahead Carlos.

Carlos L Chacon:               Travis, I think the folks who are familiar with Linux are going to be maybe a little bit earlier adopters with the docker stuff. That may not be quite a fair assessment, maybe in my lay of the land. I think that with this migration of SQL server, at least those in the data space will have to become a little bit more familiar with Docker or at least   get more familiar with it than we currently are.

Travis Wright:                    Yeah, I would agree. With Docker, to me feels like the VMs did, whatever 15 years ago or so, Right? It feels like we’re on the verge of a big transformation in how people isolate their workloads, how they provision their workloads. Docker, I think is going to be huge. We have Docker for Linux containers for SQL Server. That is from the very beginning of the SQL Server on Linux project, there has been an emphasis in making sure that we get that working really seamlessly. We’ve also been recently working on having SQL Server running in Windows containers. With Windows Server 2016 being generally available now and having the option of running Windows based containers, we want to make sure the SQL Server runs well in a Windows based container. We’ve recently pushed out a SQL Server Express Edition in a Windows container to Docker hub. People can go grab that and use it. You could take that same Docker file that’s available there, just change the location of the bits and actually have a Docker container in which that you build for other additions of SQL Server. What we need to do now as far as Docker goes is just do some more testing around having SQL Server running in a Docker container. Get to the point where we can have that be a first class supported scenario of having people running SQL Server on Windows containers as well. That’s something that we’re working on.

Carlos L Chacon:               Good deal.

Steve Stedman:                I know one of the things that I end up working on a lot is database corruption, having to fix it. Is the typical, or are the typical DBCC, check DB, check table, and those type of commands going to be included and available as well on Linux?

Travis Wright:                    Yeah, all of that works. So does tools like DMVs, those are also working. We’ve added a couple of additional DMV’s there. We have tweaked a couple of the DMVs that exposed system informational so that depending upon which system you’re running on, whether it’s Windows or Linux, we go and get the right information from the right place. Then we surface it up in the exact same results schemer.

Steve Stedman:                This is a tool that’ll only run on Linux for now. What it does is it creates a mounted file system using Fuse on Linux. It creates a bunch of virtual files, each of which represents the DMVs in SQL Server. What you can do is, you can just CD into this mounted directory basically, this fused directory. Then you can just LLS that directory, you can see all the DMVs listed there as these virtual files. Then you close as though they’re files. You can just cut them to get the upload of them or you can type to tools, like ‘cut’ if you want to use just one column or whatever. What happens is when you open and read that virtual file that we’re actually initiating a call out to the SQL Server and executing that the DMV query and bringing back the results as though they were the content of that file. That sounds mind blowing. This is what Linux administrators expect; is that they can go in, they can look at system information in a file. We’ve created this virtualized experience using virtual files where they can go and get the DMV data and view it as though it was a file.

Carlos L Chacon:               Interesting.

Steve Stedman:                That’s great. Then using your favorite shell bash, whatever it may be, you can then use those files to type them into Set, or arc, or whatever you want to use there to work in your normal Linux type environment to get the output you want.

Travis Wright:                    Yeah, exactly. Because they’re just files there effectively, you could have that in a mounted directory even and share it out through a samba share, for example. People could remote into it. It’s just a file. Right? You can do all the things you could do on that. The other thing about the DMV tool is you can actually have … You can set it up with a config file where you can have multiple SQL Servers in this config file. Then what you do is you CD into the mounted directory, you’ll see all of your SQL Servers there as directories essentially. Then you’ll see you can CD into which ever SQL Server you want to look at and look at all the DMVs for that SQL Server. Then you can do really interesting things because these are all just files. You could do Scripts over this. You could traverse over each of those SQL Servers and look at a particular DMV, compare data to each other, or whatever. It’s just files at that point. Very interesting possibilities there. I’m curious to see who will be the first person to build some monitoring tool based on this. That would an interesting thing to see.

Steve Stedman:                Yeah.

Travis Wright:                    The other thing that we’ve done is the flip side of that; where we will now have a DMV that will expose … Which will show you the process information from the proctor rectory. The proctor rectory if you go in a SQL Server … If you go into Linux is where all this system information is about all the processes running on that particular Linux host and will expose that proctor rectory as a DMV. People can just get access to the proctor rectory by running it to a SQL query. They don’t have the SS agent of the box, they can just run it to a SQL query if that’s what they prefer to do. That will be very interesting I think for any of the monitoring tool that moves out there to be able to access that information just through a T-SQL query.

Carlos L Chacon:               Exactly.

Steve Stedman:                That sounds great

Travis Wright:                    The monitoring tools and monitoring tool vendors and that kind of thing out there don’t ever have to access each into the box to figure out what’s going on. Everything can be exposed through T-SQL. Obviously is systems go down for some reason then you’re black, dark. At least for the most part it should work.

Steve Stedman:                On the Windows version of SQL server, one of the more controversial commands that’s out there is the XP_command shell.

Travis Wright:                    Yes. Yeah

Steve Stedman:                Is there a Linux equivalent of that where you’ll be able to check out your shell and run local commands?

Travis Wright:                    Right. As of right now, we’re thinking that we’ll probably close that off for Linux. We’re always interested in hearing people’s feedback on that. That’s just a security vulnerability that we’re inclined to just close off for now. There are other options for executing those types of things; like having a SQL Server Agent job for example that can execute bash scripts that will probably enable.

Steve Stedman:                Okay.

Travis Wright:                    We’ll probably have that sort of thing through the agent.

Steve Stedman:                Then with the agent today, you could create agent jobs using different stored procedures. I assume a similar thing would just work. Can you be able to create agent jobs on the Linux version?

Travis Wright:                    Yeah.

Steve Stedman:                Okay. Excellent.

Carlos L Chacon:               Travis, I have to admit, I wasn’t all that excited, I’ll use that word, about SQL Server on Linux. Now after our conversation with you, I have to say I’m a little bit more enthused about it and interested to see where things go because of, I will say, the merger, right? If you will of the 2 operating systems and what you’re having to traverse there.

Travis Wright:                    Yeah. I think we ran into that sentiment a bit at SQL pass. I think at the past summit, you’d expect to find people there that are long time SQL Server users, they’re for the most part happy with what they have today. They will maybe just want us to add more stuff to SQL Server or what they would like us to do. Or maybe improve all of your performance on scale. Whatever it might be. They live in this world of SQL Server as they know it today. Right?

Steve Stedman:                Right.

Travis Wright:                    I think as people start to learn more about SQL Server on Linux, they see the new opportunities that this presents to everybody. Whether you’re a consultant, or you’re a DBA, or you’re a developer, if you think about it it’s really just expanding opportunities for everybody in the SQL Server ecosystem. Where we can go now and talk to our friends over in the other hallway about Linux. We can figure out what makes sense to do. It’s not maybe just a one-size-fits-all solution running on Windows. We need to run something on Linux. For whatever reason, we can now. We can have that conversation. I think it presents some interesting opportunities for all of our partners in this space. Whether they’re monitoring tool vendors, backup vendors, security vendors. Presents opportunities for all of our hardware partners, we have brand new friends now with our operating system vendor partners like Suso, Nautical and Red Hat. It really just was opening up the ecosystem and presenting a lot of new opportunities to everybody. I think as we all look at the landscape and how Linux is just growing so much, it’s good for everybody that’s in the SQL Server ecosystem to take a look and see it as an opportunity for you to expand your own opportunity. Learn about Linux, expand your skill set, have more opportunity as you look out to what your next job might be. It might be to run SQL Server on Linux. There’s a … You can get huge demand for people that know how to do that.

Carlos L Chacon:               Exactly. Exactly.

Steve Stedman:                I know it’s been a couple of years since I’ve done really much of anything with Linux but I’m excited. I want to go spin up the Linux VM to get ready for whenever this is available to give it a try.

Travis Wright:                    Awesome.

Carlos L Chacon:               Should we do SQL family?

Travis Wright:                    All right, let’s SQL family.

Carlos L Chacon:               SQL family is just an opportunity for us to get to know you a little bit better. I guess I should have made sure that you had those questions we added in the invite.

Travis Wright:                    I think I saw that in the invite. I just didn’t realize it was called SQL Family.

Carlos L Chacon:               Yeah.

Travis Wright:                    Okay.

Carlos L Chacon:               It’s just our little segment that we call SQL family.

Steve Stedman:                Should we jump into it then?

Carlos L Chacon:               Yeah, let’s do it.

Steve Stedman:                All right. With technology changing so rapidly, how do you go about keeping up with technology and all the changes that are happening every day?

Travis Wright:                    Well I do a lot of reading. I read Tech Meme incessantly. I can’t stop reading it. I can’t stop reading Red It and all the cool stuff that’s going on there. I do a lot of reading. Then whatever I find that’s interesting, I go and try it, you know? I’ve never really been a developer on anything other than dot.net. Since I’ve been working on SQL Server on Linux, I’ve taught myself Ruby, Python, and Java. What else? One other language, Node. I’ve been working with node. I just go and try it. I’ve been using other database systems as well, to get what people like and don’t like about those. I’m a hands on guy. I like to actually get down and try to use this stuff and see what I can do with it. One of the funnest things I’ve done recently is one that Microsoft announced, the Back Framework back at … Earlier this year. I just sat down, stayed up all night, built out this little box that sat on top of SQL Server on Linux and got that working. That was pretty fun.

Carlos L Chacon:               Pretty cool.

Steve Stedman:                All right.

Carlos L Chacon:               Now, we’ve talked about lots of things that are happening with SQL Server, right? If you could make one additional change to SQL Server, what would it be?

Travis Wright:                    That’s a good question. I don’t know.

Carlos L Chacon:               Okay. We can … We’ll strip that question out.

Steve Stedman:                Yeah.

Travis Wright:                    Let me think about it while we go on some of the other questions. Then your editor can pick it up I guess.

Carlos L Chacon:               Yeah, if you think of an answer, we’ll come back to it.

Travis Wright:                    Okay.

Steve Stedman:                What is the best piece of career advice that you’ve ever received?

Travis Wright:                    I think probably the best career advice I ever got was from John Byce. He pulled me into his office one day, he said, “Do you know who Terrell Owens is?” I said, “I’ve heard the name. I don’t know who he is.” For those that don’t know, Terrell Owens was a wide receiver in the NFL. He was a flamboyant guy. He would pull his … I remember one time he caught a reception, ran it in the end zone, he pulled a Sharpie out of his sock and signed the football and threw it to the audience. His feedback to me was, “Don’t be like Terrell Owens. Look at it more like you’re part of the team. Make the people around you better. Don’t be such a showboat. Try to …” That that was some of the best advice I’ve ever gotten in my career. I’ve always looked back on that, appreciate it. The other thing that John Byce always preached to me as I transitioned into being more of a manager was, “Look for people that are humble, hungry, and smart.” I’ve always tried to surround myself with people that are hungry, humble, and smart.

Carlos L Chacon:               Yeah, that’s a good combination. Travis, our last question for you today. If you could have one superhero power, what would it be? Why do you want it?

Travis Wright:                    Yeah, I’ve always dreamed of having a superpower of human computer interface. Where I can just sit down to a computer, without any other input device, the computer can just mind meld with my brain. I can make computers do whatever I want without any typing or any other input just rained a computer meld. That’s what I would love.

Carlos L Chacon:               There you go.

Steve Stedman:                That would be awesome. I’ll take that too.

Carlos L Chacon:               Yeah. Once you’re done with SQL Server on Linux migration, you get the other one, the mind meld figured out Travis, come back and let’s chat.

Travis Wright:                    Yeah. It’s interesting, if you look at all the superhero characters out there in Marvel, DC, and everything, they all have these physical abilities that they have or whatever. Rarely is it like a technology thing that they leverage, yeah? I thought that was interesting.

Carlos L Chacon:               That is an interesting…

Travis Wright:                    Other than that, yeah. Very interesting.

Carlos L Chacon:               Travis, thanks so much for being on the show with us today. It’s been a blast.

Travis Wright:                    Thanks guys.

Leave a Reply

Back to top