Talking Databases, Liquibase, and Open Source /w Robert Reeves - Percona Podcast 53

by Robert Reeves, Matt Yonkovit

Link to listen and subscribe: PodBean

Listen to this episode of The HOSS Talks FOSS podcast. The HOSS at Percona, Matt Yonkovit, is joined by Robert Reeves, CTO at Liquibase. Robert and Matt chat about how everything around managing and running database has evolved and changed, and how automation is a requirement for modern database environments. Learn more about Robert’s unique background and some of the interesting things happening not only at Liquidbase but also in the open source space. We then go deep in the evolution and change into operations including the DevOps role, database reliability engineers (DBRE), and automation in general.



Robert Reeves

CTO at Liquibase

Robert Reeves is currently CTO at Liquibase Prior to co-founding Datical (now Liquibase), Robert was a Director at the Austin Technology Incubator. Robert co-founded Phurnace Software in 2005. He invented and created the flagship product, Phurnace Deliver, which provides middleware infrastructure management to multiple Fortune 500 companies.

See all talks by Robert Reeves »

Matt Yonkovit

The HOSS, Percona

Matt is currently working as the Head of Open Source Strategy (HOSS) for Percona, a leader in open source database software and services. He has over 15 years of experience in the open source industry including over 10 years of executive-level experience leading open source teams. Matt’s experience merges the technical and business aspects of the open source database experience with both a passion for hands on development and management and the leadership of building strong teams. During his time he has created or managed business units responsible for service delivery ( consulting, support, and managed services ), customer success, product management, marketing, and operations. He currently leads efforts around Percona’s OSPO, community, and developer relations efforts. He hosts the HOSS talks FOSS podcast, writes regularly, and shares his MySQL and PostgreSQL knowledge as often as possible.

See all talks by Matt Yonkovit »


Matt Yonkovit
Hi everybody, welcome to another HOSS talks FOSS. This is Matt Yonkovit. I’m the HOSS here at Percona. Today, we’ve got an extra special guest. Robert Reeves, from Liquibase. Robert, how are you doing?

Robert Reeves I’m doing really well. Thank you for having me.

Matt Yonkovit
Great to have you here. I’m really excited about today because I’ve been talking quite a bit about database design and the changing landscape of the database market. And knowing Liquibase placing that I figured that this would be a great conversation for us to delve into some interesting topics. But before we begin, I’m hoping you could give us a little bit of your background because you are the CTO at Liquibase. How does one become the CTO at Liquibase?

Robert Reeves It’s actually really easy if you’re one of the co-founders. It’s the position you do the co-founder, you don’t know what to do with. That person normally becomes CTO. So my smart co-founder, Pete, he leads product. And then, now there’s a distinction here though, because Liquibase, the company was formed in 2012. But Liquibase, the open source project was started in 2006. And that’s by Nathan Voxland. And so he is the project founder of Liquibase. He is benevolent dictator for life for Liquibase. And so yeah, it’s really easy to be CTO, once you start the company, as you just give yourself the title.

Matt Yonkovit
Well, so Okay, so how did you like come about starting the company like, like back in 2012? And did you start with Liquibase, the project before you joined the company? Or did you join the company, and then get your first exposure to the product?

Robert Reeves Well, we found Liquibase after we started the company. Okay, and so where this started is going back to the days of 2009. And so I had started another company previous to Liquibase called Phurnace with a pH. And it was a very early application server, Java application server deployment tool said before DevOps. Alright, so we started the company in 05 before that term really happened. And I am a recovering release manager. I used to call it software configuration management. Wow. What a horrible time. What a horrible title. Way back. It is so dated. And like Larry Wall, wrote in his pearl book. There are three virtues of a good software engineer, laziness, impatience, and hubris. And I have them all. And so I was really struggling. I was at this company. And we had a JHU web app. And we needed to move it to WebSphere and WebLogic. And I was maintaining two different sets of scripts. And I said this is just BS. I Hate it. Hate it, hate it, hate it. There should be an artifact that defines all the JTW goodies. I need JMS, queues, topics, all that stuff. All right, oh, JDBC stuff. And know how WebSphere likes to be talked to know how WebLogic likes to be talked to and could roll it out to both of them. And so this company was deploying the software to customers that had WebLogic or have WebSphere. And so I realized, Oh, my God, we need to fix this because again, lazy and impatient, right? And hubris that you’re very, very highly of myself. I should be working on just scripts. I should be doing more value add because I’m an amazing software configuration manager. We start a company and it worked. We end up selling it to a lot of very large companies. And one of the large companies in 09 I remember this conversation. It was the summer of 09 and I was at I had Customer Success reporting to CTO at the time other a Phurnace. And so it was our first seven-figure deal. Oh, that’s always exciting. Yeah. So I went, I got on the flight and that was my baby. Alright. Also, we had just one other CS engineer, so we didn’t have covered so I had to go

Matt Yonkovit
the pains of being a small company.

Robert Reeves And so, I went and they said, Robert, this is great. I love it. What are you doing about the database? And I said, Oh, that’s on the roadmap. It was another roadmap. The narrator says it wasn’t on the roadmap. And so I asked him, I said, Well, what’s the problem? What are you trying to solve? Let’s talk about that. And they said, Look, prior to using Phurnace prior to speeding up our application deployment, right? See where I’m going with us. We only updated WebSphere once a week. And typically, we have to have the DBAs run a SQL script to update the database. So the app will run, right, you’ve changed something in the app code that requires new buckets in the database little cubbies to put your data. Okay. Um, that was manageable when it was once a week. Then they started using Phurnace. They started deploying every day, then sometimes multiple times a day. Sound familiar?

Matt Yonkovit
Yes, yes. Continuous Integration.

Robert Reeves Deployments a great day. Yeah. Yeah, Martin and Jess, thank you both. But I really thought about it. And this was right, BMC acquired Phurnace. And so I’ve really been thinking about this and realized that DevOps was coming into its own. And I started to see the beginnings of cloud was really taking off from a corporate IT perspective. And then also, here comes this thing called microservices and containers. And all this stuff is accelerating, how quickly accelerating how fast we can get the apps out. But nothing for the database. And so Pete and I had said, Okay, we need to fix this. And he was employee number one over Phurnace. So we’re like, yeah, let’s get the band back together. Let’s do this. And while we working on it, we came across Liquibase. And we said, okay, shit. It’s open-source, it does this. Oh, my God. And we reached out to Nathan, we were like, Nathan, what do you say, buddy? What do you say we’ve got the peanut butter, you got the chocolate? Let’s do this. And I really the pitch to Nathan was, Hey, you want to work on your baby. Do you want to quit working for the man? And work on this labor of love? Because remember, open-source always starts with a developer having an itch that needs to be scratched. Absolutely, yeah. So we’re like, let’s help more people, dude. We’re going to maintain open-source. We’re going to have the Community Edition, but we’re going to add features. We’re doing open core. And so far, it’s worked. And here we are. But that’s really how I got into it. That was my first foray. I’ve always been an open-source. aficionado fan, but I never really worked on a community before Liquibase, but now Oh, man, I have drunk the Kool-Aid.

Oh! no, no, no, you’re a convert. Right. Not Now. Now. You’re here preaching the good word. So interestingly enough back in the day when you were mentioning your Java app, I’m curious. What databases were you deployed on?

Well, see you back then. This early 2000s. Right. Yeah. Yeah. And we had, we had four choices. So we had Oracle, SQL Server, and DB two. And that’s what we were deploying on because, this was a transportation app. This was an app that helped carriers figure out how to efficiently route their deliveries.

Matt Yonkovit
I might have used that because back in the day, he worked at Penske logistics, okay. I don’t know maybe I was maybe I use that app at some point or another had to support

Robert Reeves Yeah, it’s, uh it was a little tiny company, and it’s not around anymore. They end up getting acquired by their customers. But there were rumblings that this new database, this new like, thing that you don’t have to pay money for and you just get it. I’m sure Postgres was always around, but it was MySQL. And that was the beginning of it. And look, I’ve had a front-row seat to what has happened with databases, or over the past couple of decades, I guess. And back in the day, Liquibase only supported four databases. It was MySQL, and then they had Postgres. So five, early days, but I kind of stuck there for a while. And then when we started funding the community efforts in open-sourcing it, we now have 33. Wow, what?

Matt Yonkovit
Well, isn’t that indicative, though, of the database industry as a whole right now?

Robert Reeves Well, yeah. I mean, remember back in the day, right? Like, it used to be, oh, we’re an Oracle shop. We’re an IBM shop, where Microsoft shop and you use a database, and you chose, you don’t choose a database; database chose you. But now we’re using databases, based on their impact on the application? Does it make sense to use Oracle? Does it make sense to use MySQL? Does it make sense to use something like Cassandra, Mongo, or something from one of our cloud friends? We have graph databases. We have so many choices now. And now we get to choose what’s best for the application? Not based on who did the CIO play golf with most recently?

Matt Yonkovit
What did you play golf with most races? That’ll dictate a lot. I know. It’s, it’s it is interesting because I think that the paradigm shift is quite substantial over the last 10 years, when I was working at the aforementioned Penske, for instance, they were an Oracle shop that just was reluctantly using SQL Server for a few applications. Right. And you’re right, you negotiated a big contract with someone. And that was it. It was that that’s your stack. Yes, you might be a, you might be a Visual Studio shop, you might be a Java shop, you might be and you had the one database you had, or maybe two. And that was it. And now because people want to move so fast because there are so many microservices because it’s all about how quickly you can turn around and deploy. You have these developers and everybody else who’s picking the technologies that make the most sense for that particular application. And it’s caused an interesting explosion in not only the number of database technologies but the number of individual databases people have to manage and support. From an operations perspective, it’s a little like the wild wild west, where you might have hundreds or 1000s of individual databases, some small, some big some all over the place, as opposed to one monolithic database system that everybody kind of dumpster diving.

Robert Reeves Oh, hold on, let’s, let’s explore the wild wild west. I mean, if we’re going to use that as an analogy, which I think it’s a good one. All right. Let’s think about the innovation that came out of the west. What technology did we get because we as a country, we’re moving to the west? railroad,

Matt Yonkovit
Oh, I mean, telegraph, the railroad? We? Yeah, yeah.

Robert Reeves it’s we really figured out how do we our whole transportation infrastructure come from that move west? And so yes, somewhat lawless. It’s a little rough and tumbles. Indeed, there’s not certain there were amazing opera houses in Denver. It’s, it’s, it was a tent where they just had a show, as opposed to New York or Boston. And so you’re right, there are things that are lacking. But I look at this, and I say is the problem that we can’t manage? Or is the problem that we’re not allowing others to manage it?

Matt Yonkovit
Well, yeah, I mean, I think that’s a good point. Because I think there’s a combination, right. It’s not that people can’t manage it. It’s the tooling hasn’t caught up with the deployments in a lot of cases. Right. like, I think from the perspective of the massive amount of databases out there, a lot of tools are still designed for simpler infrastructure, or they might not be designed to work in a cloud-native type environment. Databases inherently haven’t been cloud-native until recently. And for some, like Postgres or MySQL, it’s kind of Bolton’s Yeah, after the fact. Whereas a lot of these new databases that are starting to pop up are quote-unquote, cloud-native, whether it’s cockroach gigabyte, they’re taking protocols that are from those databases that people love and use and trying to make it cloud-native from the very infrastructure up. So I think that there’s work to kind of evolve where we are too kind of match the application paradigm doesn’t quite match yet. But then there’s also you mentioned, like what, we let people manage it, the rise of the cloud says, Yes you look at Amazon, and how much they’ve invested in RDS. And Microsoft, and Azure and Google and their cloud SQL they’re trying to make it so it is easier for people to have those fleets. But I still think that there is a gap, because even if you’re running on an Amazon or an Azure, and you’ve got a large fleet of database servers, you still have to manage and maintain and you have to design things properly, to get-go. And I think that there’s a false sense of security, oftentimes with outsourcing or with on the cloud, where, Hey, it works. It’s good. Yeah, like I pay them to do it.

Robert Reeves Well, the old old adage, garbage in, garbage out, applies to the cloud too. You cannot just lift and shift, and not only have old architectures but old ways of doing business. In the cloud, you’ve got to change everything. And it’s not the real value of the cloud is the flexibility is, you can this is an API that you can program to, okay, cool. Let’s treat it like that. We don’t need a person sitting there caring and feeding. It’s not like back in the .com days we had a person hanging out next in sun 10k massaging Oracle. It’s just like, like, it’s like fitting it like a why to be feeding it beer and soybeans. And just like taking care of it. But Martin Fowler talked about this with microservices. He is a blog post about this article, blog post, whatever the hell it was, where he really defines what microservices are. He talks about distributed data management if we are going to allow our teams and encourage our teams to build it and run it themselves. They need to be complete masters of their own destiny. And so I think that by limiting, their they can’t touch this monolithic database, they have their own database. And whatever platform they choose, great, tear it up. I don’t care what your stack is. As long as you maintain that web services API contract, we’re good. We’re gonna do whatever you want.

Matt Yonkovit
But, you mentioned the API, kind of like that contract. SLA, if you will. It’s, it’s interesting. This is the rise of the SRE and how important it is because I think that from an API perspective, right, so you’re giving people access to your, your data, and you’re trying to give them some sort of like, run it on your own. But there’s an inherent assumption that things will perform, let’s say on par with what they would normally do. And when there are issues, whether it’s performance or outages, if there are things that aren’t working as prescribed or within norms, then somebody has to become that detective. And I think that’s the kind of the missing sauce in a lot of companies when they go fully into the cloud is that that person who, or set of people who are looking into those, those oddities, and you’re still managing things as a group. But now, instead of having one person who’s just taking care of the one thing you see this rise of the SRA type persona, where it’s like, hey, if something’s not performing within the defined parameters, I got to figure out why. And that could be configuration, it could be a new deployment that was deployed that no one told people about could be. I mean, that happens a lot, right? In fact, I went around and asked Percona customers a couple of years ago, I’m like, What’s the number one thing that like you have a problem with? Because you the most issues? They’re like, Oh, it’s code release? Yeah. They’re like, it’s code releases because we don’t know that they’re going to happen now. They just happen. Yeah. And All of a sudden, something that was working perfectly fine two days ago, is horribly broken did

Robert Reeves well, and, and it’s going to increase in, in volume and amount we have all about it as an industry to DevOps. And so DevOps shows us that change is the enemy of stability. So let’s have instead of these big, massive once a quarter releases once a year, once every half year twice a year, let’s have releases more often little tiny releases. And so little change, little instability. And so let’s do this. And let’s build the machine. And there’s got to be, we’ve got it, you got to change this mentality, that we really want a human involved in every single one of these things. Humans are really bad at rote, repetitive tasks. Oh, I’m really bad at I’m the worst. I get distracted, like a hummingbird. And so, um, but I want to focus on the more interesting creative problem-solving aspects of my job. Nobody wants to fill out TPS recovery sheets, right? Nobody wants to do that.

Matt Yonkovit
No, no, no, no, no TPS reports for me.

Robert Reeves I think we’ve got to get to this point. Where we’re just not worrying about it, we’re just realizing that, hey, this is going to happen. You can’t fight city hall. This is what is happening. So make peace with it, and figure out how you’re going to respond. And what I’ve seen the rise of recently, and my friends at Adobe are doing this. I’m so proud of them. I’m not seeing the DBA title anymore. I’m not seeing a database administrator. I’m seeing a database reliability engineer. I’m seeing DRE and oh, man, I love that. I love seeing that.

Matt Yonkovit
yeah, I think the rise of the database reliability engineer coincides with the site reliability engineer, right?

Robert Reeves Of course! The database always lags.

Matt Yonkovit
Yeah. But it’s the same with the cloud. Right? So I remember, I would spend months preparing for an upgrade with Oracle. Right? You would, you would test it, you would make sure the backups were there, you would do this, you do that. And now with a lot of cloud providers and upgrade is automatic. Yeah, right. It’s behind this time. and sometimes it works. Sometimes it doesn’t. Because we can get into a whole issue with all the automation issues we still need to solve. But let’s put that aside for a second. But now, it’s a waste of time to spend a ton of time on maintenance-type activities like backups. Yeah. did the backup work? Yes or no, that’s all I really care about. It worked. I can restore it. Okay, it just happens behind the scenes. I don’t need to check it every day. I don’t need to that’s good. I want to automate that as much as of course. But it’s those oddities like I mentioned, right? And this is, this is where the whole Rise of the observability keyword like, like, if you’re doing the buzzword cloud right now. observability is pretty, pretty big in the buzzword cloud because you’ve got so many of these databases out there. You got so many of these systems out there. It’s how do you find how does one person manage 50,000, things that are automated and then find the one or two that are causing mischief? And oftentimes, it is the small things that cause a lot of mischiefs. Not necessarily the big especially when you’re doing continuous deployment, right butterfly here. Yeah, yeah. Well, and you look at a lot of these cloud outages we’ve had over the last couple of years, and a lot of them come back to like, one configuration change. Well, sure. Right. Like it’s just one little tiny thing. Oops, I changed the subnet or something crazy, and, like, everybody went offline. Yeah, it’s always yes. It’s always saying yes, yes. Yes. It’s always it’s always guaranteed. No, yeah. Or Cloudflare. Right. So, like cash.

Robert Reeves Remember going back to agile manifesto, and updating how to freak in old I am. But going back to Agile Manifesto, we also saw the first implementation of that, that I saw was extreme programming, and so early 2000s And so all the books the x was capitalized, this is back when everything was extremely I think we’re relisted a lot of rap metal, extreme Doritos extreme.

Matt Yonkovit
It’s like we go through this society.

Robert Reeves Despite its wacky naming, extreme programming, I really like what their approach was I really forget the original, the originator of the idea, and I apologize for that. But the theory was, if testing is good, let’s do it all the time. If code reviews are good, let’s do it all the time. And so that was pair programming with extreme programming. If testing Well this was around the same time as test-driven development, before I write code, I’m going to write a test to tell me if I’ve done the code correctly. And that’s actually how we do it at Liquibase. When we expand other databases, how we were able to quickly validate support for the Percona Xtradb cluster because we had all the test cases. And thank, thank you, Percona, for building it on top of MySQL, we were able to leverage that. But having these tests having these code reviews if continuous integration is good, let’s do it all the time, Martin Fowler started out with CI. And so I think, if releasing is good, we should do it all the time. If automation is good, we should automate all the things until we need a human because you’re always going to need your Columbo in SRE DRE land to figure out what happened. And so instead of fighting it, I really would like to see more companies embrace it, I would really like just and what, it’s not at the technical side of the house. It’s certainly not the newer younger folks’ handful of years experience, they’ve bought it. It’s no management up with it. Yeah, it’s leadership. And so I really would like to see them say, Hey, man, look, let’s do this. Instead of just saying, well, we’ve got to do it, how we did it. In 2003, like, come on, really, really?

Matt Yonkovit
Well! The good news is eventually like it’s all cyclical. So today’s new programmers will eventually become tomorrow’s managers. How we can accelerate the process, though, I think, is the question, right? Because I think that there is a lot of value out there. And the good news is, I’ve seen a lot of companies start to like, develop their own new teams around whether it’s the philosophy of automation or DevOps means it is very trendy right now. And I think that the laggards are starting to come around, especially because you start to see analysts talk about it. Right. Yeah. I know that there are some companies who don’t move without Gartner telling them where they should move to. But so it is a little bit laggy. And so there’s a lot of let’s be honest, DevOps is not new anymore. It’s established, it’s there, you should be using it, you should be using automation. But a lot of us are looking at the next thing or the next day, right. Yeah. I think it’s, it’s important to know that it is, is out there. And right now, I gotta say, I don’t know of any new application that isn’t being designed with that automation in mind with that kind of cloud-native microservice architecture. It just, I don’t think people are developing new applications that way. If they are it’s very well it

Robert Reeves Like I said, cyclical there was a time where it was web first, then it was mobile-first. And it was cloud-first. And now, so companies learn. But the thing that bugs me is that well, it’s William Gibson, right? He said it. The future is already here. It’s just not evenly distributed. And what bugs me? Yeah, what bugs me is that there are these teams are are these pockets of greatness in these larger companies. But there are still people on. And this is not a slur on the legacy apps. Meaning the legacy is a good thing. leaving a legacy is a human Carnegie that that’s a good thing. But if you are Carnegie, however, you pronounce it, whatever. But those legacy apps throw off so much revenue billions of dollars for these companies. And it pains me that they’re just nibbling around the edge on the newer apps. And not figuring out how now there are some companies that have figured it out. There are some companies that figured out how do I take this old busted JHU web app. And as I’m continuing to improve it, I am refactoring it on the fly and peeling off functionality in microservices, I’m using the Expand contract, expand contract migrate pattern to add microservices, and move people out that that’s cool, man, that that’s good, but not everybody’s doing it. And I’m just thinking about all of the happy hours, birthday parties piano recitals, fishing that’s being missed because somebody is dealing with this stupid crap. It’s terrible. But if job security that’s the problem! Hey man, it’s 2022. Nobody’s worried about job security right now.

Matt Yonkovit
true story. Okay, true story. I actually worked with a company, not too long ago, big company, a fortune 500 company, where they wanted to modernize their internal, IT systems. modernization is a big topic, right? modernize the new CIO, what’s the modernize? Great! Well, and this is where politics, and not the politics of the country, government dictate so much at a lot of companies, this is a people problem. This is a people problem. This company is brought in new management. And they’re like, we’re gonna go all-in on open source databases, we’re gonna go all-in on this new stack, this new framework. And so they actually put financial incentives on the management team to hit certain goals. We got this many applications migrated, we went this many Oracle, to MySQL, migrations, all this stuff, because they wanted to modernize. And they wanted to make it more cost-effective and more efficient. And the teams that supported the legacy applications, they, they, how the teams were structured was this application, this business unit was in this face, they’re like, We have no incentive because their management was different than the structure management team. And so they just said, No. And a year later, that entire management team is gone. And someone new came in with a brand new stack and a new idea. You’re ready. And it’s one of those. One of those weird things that a lot of our choices and decisions are dictated by like you said, the team and management, the team and power that the new people coming in, who have some sort of decision making, and they can change or attempt to change a cultural overnight. But if they don’t have the buy-in of the development teams, and of those actually supporting them. It’s an uphill battle.

Robert Reeves This is how I would have approached the problem. Alright. And this is where and this is my idea. I’ve seen this success, successfully done it, many, many companies are to not change the stack. And to say, you have to offer services, and you have to guarantee the performance of those services. And so instead of this app is not a microcosm. You have a connection to another application. Right? So maybe let’s imagine it’s a bank, and you’ve got the A team that does investments, Wealth Management, whatever. But they need to connect to the single sign-on. And so the single sign-on folks are going to offer their identity management to the wealth management, folks. Okay, cool. All right, here it is here’s the API, this is what we’re using. But at the same time, I want to be able to put little widgets on the main homepage, which is the consumer side. And I need those widgets of account information, those sorts of things from the wealth management folks from the investment folks are, you’re going to offer that to me as a service. And that’s, that’s going to happen, make peace with it, I don’t care how you do it. And I would have put incentives on them to do it. And then once it started happening, I would measure the response time and rate of the performance of the application of the services they pushed out and constantly seek to improve them. So last quarter, you were x, we’re trying to get better than the access quarter. Okay. And then also measure them on what Dr. Forrester and Dr. Nicole Forsgren. She’s done a huge amount of research, she wrote accelerate with a bunch of other folks. And, there are four key DevOps metrics. And that’s what I would hold them accountable to, I would hold them accountable to a number of releases to production. And I care about dev and test just production, number of releases. The success of those releases the meantime to recovery. So when something bad happens, how quickly do you get back. And then, oh, my God, I cannot believe I’m blanking on the fourth, I’m having a Rick Perry moment. But having those four metrics, and then also laying out it sounds like in the example that you cited, that incentives weren’t aligned. And when you’re not incenting people, you’ve got a whole Who Moved My Cheese issue, and you can’t fight that. Like, like, you might win against City Hall, you ain’t winning against that.

Matt Yonkovit
Yeah, yeah. Yeah. And it’s, it’s a challenge for a lot of larger companies that have those politics at play, even smaller companies end up with politics. And I mean, I’ve seen this numerous times. And one of the problems that plague the open-source space is a lot of companies in the open-source space, provide services as a primary vehicle for their revenue streams, right. So they’ll offer support or manage servers, consulting, those types of things. And in the case of teams, a lot of them will say, who I want to support, because I value it, or I want the enterprise version, if it’s open core, because I value it. And then someone changes from like they bring in some new DBA or new whoever, and they’re like, I would never pay for that, because it’s open-source, and then they end up migrating off of the enterprise version to a community version. And then a year later, they leave and someone else comes in and goes, Oh, my God, why aren’t we using the enterprise version? And so I’ve seen this kind of ping pong back and forth. And I think a lot of companies have that mentality, and they ended up just not being able to move forward. So I think it’s a big problem, but it’s huge. I agree with this function. The very definition of it. Oh, I agree with you.

Robert Reeves Absolutely it just hit me the last DevOps metric was lead time. So when the developer checks in code, and then it’s on production, sorry.

Matt Yonkovit
Yeah. Yeah. Yep. Absolutely. And I think all of this ends up getting tied back to the database because we mentioned like, the database is so foundational for all of this. And, honestly, it’s not all that. For a lot of developers, it’s not interesting. It’s not as interesting. So we don’t see the innovation from them on the database side. a lot of these companies that are designing these new databases are picking up the mantle for some of moving the ball forward. But there are a lot more developers working on app code, who just don’t want to think about the database as much. They just want to plug it in and let it ride. Yeah. This means the automation really speaks to them if they can just use it as a service. If they don’t need to think about it, it just happens. That’s awesome. Yeah,

Robert Reeves Good. Look. I want developers to focus on solving really hard business problems. And if we can make it easy for them, to get their changes out to make it a non-issue on the planet form, it’s always going to be available. That is a good thing. Um, there was a time we say, well, not everybody, maybe people my age, say refer to things that always have to be up as dial tone. It’s an old dude saying, but they’re those people who don’t even remember dial. Right. But scary. Yeah. For the kids that are listening, there was a time when you would pick up a phone, and it would make a sound saying that was ready to be perceived numbers to be dialed. There was a time when dial tone was not expected. It’s like, oh, the phones work. And that’s a good thing. Because you had people involved, you had this, this switch operators, were Yeah, you would call up and somebody would manually connect your call. And especially if you were doing long distance, you would have to go LA to Washington DC, you would actually have somebody would have to plug in wires to connect this. And, and it was a mess, and it was a pain. But then we got automation, we got automatic switching, which totally changed how we communicate as a country. We disintermediated the switch operators. Now switch operators, they were an incentive, right, I don’t want this automation, I’m gonna lose my job. Yeah. But then they go, and they wind up being the people that installed the switches, you start seeing the rise of the telco engineer, instead of somebody who was just manually plugging in and pulling out wires, they’re actually going and installing the stuff. So we start seeing an increase in value add for that telco employee, and also how much they’re making. They wind up learning more, they’re more valuable to the organization, but they’re also making more money. There is an incentive to learn new technology and automation. So this by humans doesn’t like it. The individual habits have a real hard time saying, wait a minute, which impacts me negatively. And as a species, we tend to veer to the negative. What’s the worst that could happen? And now really focus on Well, what is most likely to happen? And is that good?

Matt Yonkovit
Does it remind me of a commercial? I don’t know. I don’t know if you’ve ever eaten at Quiznos subs. I’m familiar with the place and what they d? There used to be this commercial long ago, I think you can find it on YouTube. And they said hey, they had the first hosted subs. And there were these kinds of Neanderthal people wandering around. And it’s like, it’s like the first guy who invented pants. Yeah. And they all have bushings in front of them, and behind them, and like, this guy walks out with khakis. And what are those that like their pants? And one guy goes, I fear change. I will keep my bush and he runs away. Yeah. me, that’s kind of what it is. Right? It reminds me of that. Yes.

Robert Reeves And that’s that is never going to change whether it was the move from mainframe terminals to client-server, client-server to a web moving from on-prem enterprise software to SaaS. These were big moves and the same conversation that we’re having was had back then. Except he was what? In the Letters to the Editor section of PC mag. Podcasts. He didn’t have that. Yeah, that’s right. Yeah. It was on us that. Yeah. Water Cooler discussions, you know? Yeah. bulletin boards back then. Yeah. But here’s the thing. I’m hoping that we have learned and we’re starting to accelerate. And I think the key to that is open source at the end of the day. That is where that that marketplace of ideas has proven out to be far, far more valuable than anything, an individual or an individual company could come up with

Matt Yonkovit
In crowdsourcing, really and you mentioned being lazy. I don’t look at it as lazy. You just want to do you want to find a better way. And from an open-source perspective, you see a problem. And this is part of the power of a lot of the projects that we work with. You see a problem and it doesn’t quite solve it. You want to go fix it. So you don’t have that problem.

Robert Reeves But that’s hubris. That’s the pride part of it.

Matt Yonkovit
And I mean, but and I think that’s the interesting thing, it’s a different paradigm shift for a lot of developers because they’re used to, I have a goal in mind, I have a target. And then if I hit my target, my boss is happy and great, but they’re not really invested. This routes investment four people into the project, because it’s like, yeah, I really don’t want to spend my time doing this. Right. I’d rather just have it automatically happen behind the scenes.

Robert Reeves Well, yeah, yeah, yes, check out three And because that is where they have those three virtues from Larry Wall, and they describe it, it’s just one page on the website. And it is it’s phenomenal. It’s so good. And I just really wish we would get faster as an industry of recognizing that. So for us, it Liquibase when we look at the community, we for we view community requests as an indicator of areas that we need to beef up on the paid offering. So when people start saying, well, things like, Okay, well, I want to be able to have Liquibase reach out to vault, or CyberArk to get my authentication information, username, and password. Um, what we did is that we added an extension framework. And so we first enabled other people to do this, like, Okay, you want to do this, here is a way of doing that, instead of getting into the code. And hacking the monolith, the core, well just add this extension. And then what we’ve done is based on that feedback, we have added extensions to our paid offering. And so we’re enabling people to be able to do it themselves, we’re never going to stop people, we’re never going to on open-source things or change our license model. We’re Apache 2.0 that is free and you can get it. And we are going to we’re going to look at that stuff and say, for our larger customers, we’re going to help them out and they’re going to pay for that. That assistance, we’re providing them, but we look at that community as a way of telling us where’s the puck going? Where do we need to skate? As opposed to lagging indicators. And I think that right, there is why the community is so valuable. It’s not just free code, that’s lame. What’s really valuable about it is that you get these ideas for people that might not be paying customers, and just have a problem. And that sharing that information, that information, that little nugget, that morsel, so valuable.

Matt Yonkovit
So that’s a big mistake, a lot of open-source projects and companies make it they attribute contributors only to code and the most valuable contributions or feedback, and ideas.

Robert Reeves Evangelism? Oh, my God, I just use this open-source project, and it’s so good. And I got to that baseball game with my kid. Yeah. that is more valuable, I would argue because they tell one person, the other person tells two, so on and so forth. You got exponential growth. And remember this the platonic ideal of this is the Linux kernel. And you see, Linux, just reaching out on that Minix subgroup on Usenet, back in the day, hey, I’m working on this thing. What do you think? And start getting people to provide feedback? It wasn’t the code. He wanted the feedback.

Matt Yonkovit
Yeah. And I think that’s the that’s something that I think a lot of people don’t quite understand about it. But, Robert, I hope that we can get you out to Percona live this year. I know it’s right there in Austin. So it’s just down the street. It’s right down. Yo, yo, come out and see us. But I wanted to thank you for coming on out talking to us a little bit about DevOps automation, a little bit about Liquibase, all things databases. Appreciate the time here. It was great catching up.

Robert Reeves Well, of course, and look, I the answer. Excuse me. The answer is yes. I’ll be a Percona live and if anybody wants to chat about this. I’ll be around and it’s really easy to find me. You can blow me up on Twitter at Robert Reeves. Find us at Liquibase and we’ve got a ton of good stuff out there to help our database brethren in the trenches. And the stuff is free, it’s open-source. You don’t have to pay us a dime, it just breaks my heart to know that people are still doing this manually. So we’ll help for free too on our discord chat,

Matt Yonkovit
So check out Robert over at Liquibase and check out the open-source project that they’ve got going on. And maybe there are some awesome things that you didn’t know you could do that you might be able to do if you check it out. So, Robert, thanks again, and appreciate you hanging out.

Robert Reeves Thank you. ∎

Did you like this post? Why not read more?

✎ Edit this page on GitHub