An introduction to OpenShift.io, an end-to-end OpenShift development platform in the cloud

if people want to get settled I'll say thank you very much for for showing up given that it was a secret session on the agenda and now we got a full room so I'm pretty pleased about that I'm Todd Mancini product manager for developer tools I've joined on the stage by Pete Muir who's one of our lead technical directors in our developer tools and I didn't know its title so I just call the developer guru James Strachan who will be assisting us today as well so thank you very much for you know the pleasure of your time this morning you will note I am sporting this beautiful OpenShift i/o t-shirt on the on the way out you will get a card and if you bring that down to the expo floor to our developer area I figure what the name of it called the devs okay here's good 7th bring going Deb zone you can get one of these shirts with that card and you get all sorts of other swag if you go on some quests if you will so much you read will observe be able to explain more of that later but I assume people are here to learn more about open chef chai oh yes yes that's so very great all right so hopefully people were interested in the announcement let me what I'm going to do is I'll just do a little bit of talking I'll let Pete and Jane kind of really show this stuff so you guys can really get a feel for it but what I wanted to do is first of all just remind people that right that does have this awesome developer program that we've been building up for about three or so years now definitely join free to join just go to developers on RedHat comm and you know you can for example download all this stuff me whatever one write if you're like wow I want a copy of rel so I could do some development on Red Hat Enterprise Linux you want our clips tooling whatever it may be get all that stuff there it was a great great content we're constantly pushing new content up there one things that's been super popular are our cheat sheets and our quick starts so all all available to anyone that that registers in developer programs so but I want to tell you about Obito and how we got to this place why did we even embark on this journey of creating open chef do and I don't think anyone in this room is going to be surprised by the fact that people on stage these days tend to like say things like that every software every company is a software company but what's really interesting is if you actually go and look at what was voted as the top 30 innovations of the past 30 years a full 20 of them are directly related to computers and software so it is absolutely the driving force of change and competition and there's this just a demand of all of our customers that we are able to innovate faster we need to compete more we need to do that by adopting practices like continuous delivery right people wouldn't have their smartphones and they're used to their uber app or whatever it may be being updated you know several times a week maybe they expect out of all applications whether they're the oldest of the you know enterprise type applications and the reality is time to market becomes so important those that can move fast win you know so all of you in this room that are developers you are the engine of competition and we see this in the way organizations are spending their money massive decrease in kind of traditional product development and product manufacturing massive increases in just allocating funding to software creation and service offerings and what we look where talent is being hired everyone I think here knows it's not really a challenge to be employed in this industry massive hiring for people with programming capabilities data scientists are probably one to grow with the fastest growing ones right now so but what we're finding is great we know how to write software so what's the problem we'll just keep doing that right but the way we build software most of us is really starting to fail that say we have so much weight and so much process and it's really starting to have these develop processes crumble underneath themselves I've people have said I'm attributed with having said that if our industry was the construction industry we'd all get all these houses built and then days later they'd all be falling down right the there's this I learned about this city in China that apparently that kind of happened and they just people all these like unqualified contractors within they erected this huge city and now it's like this abandoned wasteland of falling down buildings see we tell you what the challenges are when we talk about where do typically creep in we see a lot of issues where people have production environments that are different than their development environments massive environmental changes that lack of fidelity leads to those bugs that you only find in production and we know how expensive those can be we also know that everyone's trying to move to this world of these orchestrated micro services I mean micro services are really a great way to achieve continuous delivery because you get such awesome isolation with the different parts of your system you don't have to deploy this one massive monolith where you touch it and poke it in one area and it breaks in another getting that that degree of isolation so important to micro services really do make that achievable the other thing we've seen though is at the end of the day we have to be honest with ourselves everything we do pretty much day to day is just guessing right we just guess we say oh let's go make this change to this offer because we think it's going to do something good right and the worst thing is oftentimes the guesses that we make we don't verify them what do we do we we say hey let's go change the icon for the shopping cart to be a bigger icon because we think more people then complete a shopping transaction when they go to do a checkout we say alright that's great let's go do that right and then what happens is we then write the like go make the icon bigger and what's the acceptance test for that the icon is bigger right which kind of lost track of the fact that the acceptance test should have been that more shoppers completed a shopping transaction so we've got this massive disconnect between what the business is saying the software changes should be and why they should be these are all our guesses and then how we prove whether or not we had success and be you know we high-five each other when the icon is bigger on screen but maybe actually we have fewer customers checking out because it's really annoying to see that massively big icon we've got to get past having these guesses so some scary statistics I think people here are aware of these but some of the I find incredibly interesting now we've got the kind of standard chaos type statistics still doesn't change 71 percent of projects today are still considered challenged or failed by the people that worked on them and then when they're finally completed although we have noticed in the past ten years or so that agile software projects do have a 4-time success rate over traditional waterfall project so agile is working now that's not saying that scrum is working or XP is working it's really talking more about truly agile methods and I think continuous delivery as one of them is super-important we talk about continuous delivery so I ask customers and like are you going to confuse delivery and 50% of them say yes we are we're going there soon I'm like oh can I see your basic plan for doing so and only 12% of them can even produce a plan so everyone says they're doing it but no one has a way to do it developers come onto projects they spend half their time just understanding what the software is what the project does their managers say we got to go to the cloud that more than half my team doesn't understand how to do cloud-based development and for me one of the scariest statistics and one of the reasons though this is kind of the first thing I highlighted this morning during the keynote is only 23% of developers today when asked are you confident in the software stack choices you've made for the project only 23% of them say yes more than three-quarters of them are like unconfident about what they're doing defects are abundant 750 defects and every 50,000 lines of Java code for example those are bugs that get to production and when you have that bug in production what does it cost you well it turns out it costs you three hundred and thirty six thousand dollars per hour and how long does it take to fix one of those bugs takes about 200 minutes right so that's what two and a half three hours that's not good right that's if many of you are deploying Java apps that are what who has a 50,000 line Java app right who has a million line Java app right si some of you probably have 10 million on Java apps these numbers get big and scary real real quickly and then the last thing you know developers like wow it's really hard for me to get the stuff I need I needed I need equipment I need an environment to run in you know 75% of the developers say it's that's really challenging and so they make these requests of procurement and finally the shows up and it's not the right thing or it comes really late so what happens is if you go across organizations especially in enterprises the dev and test infrastructure they've set up is only utilized about 10% 90% of that hardware is just sitting idle burning electricity being bad to the planet you know it's like we can't do something about this so we want people to think so we we thought about those challenges and we said what would a new and better world look like and we had these very interesting meetings where we came up with interesting project names which is a story for another day but we decided if people could just move to this kind of end and integrated cloud native development environment it was just there that they could use it was instantly available that would solve massive amount of problems then we said we've got to make it so it's super easy for people to do deployment pipelines that include doing things like integration testing doing stage deployments doing a clean rolling deployment when you're finally going out to production safely migrating transactions from old versions of software to the new and then we hired a bunch of data scientists and we said let's start looking at code let's start analyzing code we have a ton of code internally at Red Hat obviously it's actually not internal in Red Hat because it's all open source so you all likes us to it too and then we look at all the other open source projects out there ones way we don't even use we start analyzing them like what information can we extract from that it's an amazingly interesting big data problem and how do we start applying that to the code you guys are working on today how do we make that become a reality and that led to open shipped IO it's really an essence talks about doing this virtuous cycle and we kind of just view that virtuous cycle in three different phases and analysis phase a planning phase in a creation phase and it just repeats so let's analysis means that means you you look at the world around you it might be here's a business problem to solve it may be hey that server over there is running at 99% CPU capacity right they can their their technical measures their business measures there's all kinds of measures we can do but it's taken those inputs you look at them you come up with some hypotheses so rather than guessing let's use a nicer word let's say hypothesis where a hypothesis implies you have a way to verify it right so maybe you said hey we don't we don't we'll like the shopping conversions we're getting so we're going to make the shopping cart icon bigger and the hypothesis is that in doing so we will increase our sales conversion rate take that put that in a plan sign that work out to people have them create stuff now creation today for us starts with the developer but we want the future of OpenShift to which FTL to include testers and designers and business analysts and every other person that has stuff that they need to create as part of something and what do you do you take that you have to create it you get it deployed and running and then you'll look at it again you analyze it again you just keep doing that over and over so that's my preamble so why don't I turn it over to Pete will jump into the the product so this morning because timing was so critically we kind of what the the video route which funny enough failed at one point oh such thing it's unclear what's safer but this is a bijective i oh here it is absolutely running in Pete's browser how could we prove that and you can click on it some and and where we should we go first should we create a new project would that be interesting you want to go into a project how about we create a new space so please create a new space he just gives it a name and we kind of went really quickly over this notion of the template so the template really is about configuring your space in a meaningful way today we'll be honest that that configuration is fairly light but we have a grand vision for how templates can really augment the way a space is set up and runs but you can see Pete has four different choices here agile scrum issue tracking and scenario driven planning so the first three probably sound familiar to you like everyone kind of knows with agile is generically people know it's from is issue tracker is kind of like what you get in a github type world it's like just endless lists of issues it's very urban planning is interesting we'll take a moment on this one this is actually the model we use in our group so when we go and do planning and development we we follow scenario driven planning so what is it so it's an urban planning takes an extremely hyper customer-focused view of the world it is not about planning features if all you do is plan features you get massively vote loaded software and a lot of organizations do this today they like let me take this feature and in the next version we got to make that feature a bit better and all the other features better and keep them more and more features but in scenario different planning you start with a scenario a real-world problem the customer has and you break it down into the value propositions that the customer will realize if the scenario is met and you break it also down as the experiences so imagine the experience like you're writing the demo script for how you would show that that scenario being resolved in your product you're writing the demo script first right because oftentimes if you did that first you would come up with a much better user experience than when we kind of build a lot of technology and they're like oh how are you gonna string that all together and so that's kind of the essence of scenario based plan so go ahead and create the space and now it's created and that's it it's done he could he could start using it right away but he's got a few kind of little helpers that he can call upon and the one we like the most is to select a technology stacking pipeline so it takes a moment cuz it's now there's a lot of work going on behind the scenes to analyze what's there and it's parsing out a bunch of data to figure out all the available application specs and now he's got all these application stacks available you can see vertex is the fault one there but these are various runtimes you know a lot of people are interested in spring boo you know you can do spring foods we're going to add many many more here now some new that are super suit or probably like hey those all look like Java runtimes and they are it's fair to say you know our primary expertise has been in the Java world most of our customers have been Java customers but that is not what open ship that IO is about that was just what we have today we are very quickly on a path to include dotnet JavaScript for nodejs type applications and probably likely Python after that if we had to get our gokumon are doing a session straight after this one to show you a little bit more about che which i think toad's will talk about in a minute and kokum is going to do a HAF example with che in that fantastic so go ahead and we'll do vertex for now why not so there's another announcement that kind of was out today about the Red Hat open shift application runtimes and I get a credit you know point every time I say that's a mouthful or roar and say like to say and so vertex is one of those applications factor you can use there from R or and lastly pipelines you know James Dragon has been so key in bringing a lot of the experiences they have on the fabric aid project to the temple you've heard of fabricate is kind of essentially the upstream and the ancestral father of openshift IO and one of the great things about the way the fabricate team has worked there Sir James will talk about this is their adoption of these continuous practices and right now Pete has basically extracted all the knowledge that James has by just clicking a radio button and going on from there why is there why I'd even know next on screen what is next so a this product it's continuous delivery so it even has new features when I look at it from 1-3 days ago so apparently there's another step that I'd even know that was in here so let's all see it currently it's the loading next step step all the organizational data so this is about getting your github repository together so the one of these ways we got to where we are today on the day we did we were you know we wanted to make sure we got launched on today its summit so OB Jeff Jo does not today include its own git repository it relies upon github so if you have a github account great you can already start using it and we believe there's actually a really good marriage there between what open Jeff Jo delivers and what github has and we will augment that with other capabilities and error messages but we will we'll add an other integrations for other get providers including your own in the in the near future so all right so this is why we've never gone to prepare before because it clearly does something wrong okay I was a different name the name here and somewhere there's someone furiously redeploying pods then open shipped somewhere right now setting up a WAP tokens so this is going to what we'll get out of it yet James likes what he sees you think it's gonna matter so once this is done then what we'll do is we'll we'll go into that same trade insights application with show day so this one was basically build that same vertex thing so this is the analyzed page and once again in this first kind of developer preview admittedly it's fairly liked it's a great way for you to quickly get to come of the hot items in your space work that's been assigned to you things like that but we really want to start bringing in massive amounts of data into this like real-time monitoring things like that that's that's to come but you will see that that stack reports that kind of gives you some deep insights of potential issues in your currently deployed code is avail maybe worth of you since that's the part that failed in the video today so it's like this is showing us at this bill that Pete rolled out actually has a security vulnerability in it a bunch of dependencies I think one or two of these metrics is probably not real yet I don't think there was a thousand 24 files in your sample app their work we're going to have a discussion on the definition of micro service perhaps doesn't sound right but it also shows you like hey here are some of the dependencies or in the application and recommended versions to upgrade to hmm all that data is real there's a thousand 24 files in that I'll including all the dependencies that's what it is okay interesting all right well it just showed that you bring in like one little innocent left pad dependency and the next thing you know your application has 147 thousand lines of code great number of files is there on the right look at that fantastic so there yes it's the log back that's the one that I added and that's the one that brought in 345 and that's the ones going to see here it's gone the CV so I'm a product manager I once was a software developer I'm not anymore so that's why I'm responsible ah great [Music] so let's get out of here so should we go to planning so in planning this is where you you do have this backlog if you go through here can you scroll down a bit it is at some point hierarchical in nature I think your agent doesn't teach me that was not so you can do a hierarchical breakdown so if you have that scenario and you want to break it down by value props and experiences then you break those down usually by features and furthermore on by development tasks and things like that you certainly can do that in here the task board is the view I like better I just feel it's a more natural way of kind of doing agile planning great for stand-ups if the video was working better you know I love using a touch screen type tablet for these you can just drag it with your finger which is really cool and but you know this is a great way to assign work out kind of get details on things link things together to indicate dependencies between one task another the other kinds of things you expect in some agile planning tools and then and you and if you come to the che talk which is the one immediately after this one will go much deeper into what's going on in create I mean this is where really the kind of rubber hits the road and we think this is one of the real key differentiators is the fact that you can so quickly create one of these environments like I said it I said it takes under a minute I think the environment that I made for the demonstration took about 33 seconds to create and I will tell you that as you know someone who has a rich history with a lot of IDs I've really been amazed at how much I like using Eclipse Chang if you had told me three years ago you're going to actually enjoy using an IDE in a browser I really would have not thought that possible and I would have felt like I was going to be so limited what I can do but there's two important things here first of all che as an ID is actually really good it works well it's fast even though you know your remoted into your this cloud-based environment the speed is great but the other thing is because che has needs workspaces this development environments been built for you the fact that that gets everything set up for you automatically and correctly and then everything just runs there that buys you so much that a lot of the things that you needed to do in those more traditional desktop IDs just goes away if you don't have the same kind of problems so pizza in here he's changing code he's probably going to put in some inappropriateness oh yeah you got so you can see you get full autocomplete and all that you'll so you'll see all of that in the che talk unit testings in there and the fact that you can just run the application directly in here now the the other thing that's is christianism so you can build if you didn't install the run command but he can but but it doesn't matter they didn't install the run command because unlike some other environments where they try to like protect you and I like I mean I love the fact that this makes it super easy for even a fairly intermediate level developer to be productive what stinks is when your code doesn't build but what really thinks is when if you're a more of a power user you're you're denied the opportunity to do the things you want to do so as Pete goes and fixes his weird error so that's about is pick a string yeah but Pete can actually go to a terminal window and he can just run maven commands you know look at his whole containerized environment do RM /rf whatever i put of her or floats your boat and and everything is just there it runs and operates superfast and that's because behind the scenes we are making use of open shipped online and it's a tremendous environment in which to do work and Keith is now doing a maven compile and vertex run and sure enough the look at that he actually got it running really good and so now he can hit it with a curl command you know so and if something's not installed you like you like the curl command what's it there you could yum install things and the maybe not today the and the data came back so that's all good so chase looking great but we're getting close on time and then the the build pipelines are actually pretty impressive so what you see here is bean already had one he goes to the full view that one that he just did X but as a coach 8 that's going to be queued up to to do a future build but one thing we don't hide anything from you so he can go into a view log and we have SSO with our Jenkins back-end system so it is a Jenkins based build process full access to the administrative console there to go look at the current run see how it's behaving get all the build logs till the prior runs all that so if you want to go into the 1997's UI that is Jenkins you you're free to do so but if you don't want to go there you don't have to and you can just say in open ships at i/o and he can do a promotion and all that well he just says well this is all you know this is all controlled by Jenkins files that sit inside your source code so you know we're not doing magic here we're not creating these pipelines for you if you want I mean these these pipelines definitions are carefully crafted by us to to work well but they're just groovy files you can just change them add stages remove stages do different things with them so this is not you know we're not locking you into that one setup this is did you say groovy this is creepy I'm speaking of every file why don't we invite James on the stage you know the little thing or two about groovy I'm sure and thank you one thing to mention by the way about this demo is everything's running on open shift right so often developers have like Windows laptops or whatever and you test on Windows and deploy on Linux the important thing to remember on this is everything's running on open ship is container so you're developing on the same environments production right you're developing on containers everything you're seeing on screen is in the container so you're developing and testing and running on the same platform all the time which saves huge amount of time right never test on Windows never test on the Mac testing then it's content all the time I'm just going to say before we came in why my circuits just won't connect to the video in here so rather than panicking as you normally do when you can't do a demos that you can't get your laptop to work I just logged into the web browser over there all I need is my password and part of the fact I can't use the keyboard and the mouse isn't what I'm used to it's all fine and Weather Service's I had to say that I did a demo last week and I ran it off my phone plugged into a docking station just use Chrome on here and it worked fine so we can show that if people want to see that later so I guess I should talk about upstream should Moline yes so so everything you've seen on the screen and everything on open shifter IO is all open source right where Red Hat so everything we do is open source open core did you say open so oh okay although core is a resources with no open code we are not open cut we're closed though we would know on the open code company so everything's open source right and what Red Hat does is we have a strong separation between upstream communities and products so open shifter IO is the product the free platform for developing the code behind over chip the IO is all in an open source project called fabric 8 fabric with an 8 so fabric 8.0 is the website fabricate as it sounds like to fabricate to make things right so fabricate is the extreme open source project now it's been around for maybe four years now it actually started before docker if you can imagine such a thing and we actually use fabricate to fabricate fabric it so we use fabric it to build fabricate itself right we started fabricate with just one get reaper so it started off as a monolith with a single release process right and then over time we started coupling it and making more and more separate micro services when it got to about eight different micro services it used to take meal or another person on my team an entire day to do a release because you'd release the first git repository and get a new version then you go to the next git repository and update the poem or the packages and with a new version then you wait for that to release some stage and then you do the next one or the next one the next one you take an entire day of watching scripts and watch at tiring logs and all this kind of crap and now we're up to about 250 git repositories right so you can imagine doing a release of 250 different micro services that would take quite a lot of time if it was manual so what we ended up doing with public 8 is building confuse delivery pipelines to automate all of this stuff right so now we can wear with a relatively small team were able to release 250 different microservices all the time right now let me just describe that a little bit it's kind of cool what we what we started off doing was we'd release the first thing and then when you releasing the second project it would look to see is there a new version of my dependencies and if so update them on the fly right so each project were the datas dependencies as it was about to being released what we found there was you'd do a change upstream and then no one else would notice you've broken something until you just have up to release which is normally the time when you're trying to get a new feature out so we switch to doing much better a job where now whenever we do an upstream release that release pipeline then automatically generates pull requests against all the downstream projects so the first project to release it will as part of the pipeline it submits pull requests against all the other projects when you submit a pull request on the downstream projects that then generates continuous integration pipelines we send runs all the tests on the downstream project if that works it then merges depending on whether you've got human approval or not and then that kicks on then a CD pipeline event downstream so automatically we make one code change upstream and everything just automatically goes through continuous integration and consists delivery all the way down so we can fabricate automate all of our continuous integration and continuous delivery of all of our software including fabricator IO another thing which is really really cool is openshift I always changing pretty quickly right we're developing code really quickly and getting features out live whenever we make a curl change in the UI for example there's I think maybe six or seven different git repos with different NPM modules and whatnot the makeup the UI if ever you make a coach in gin any of the UI modules it regenerates a new docker container for that pull request and stages the UI and in github you get a pull request comment with the URL to test out the UI so every time there's a UI change you get to visually try out the UI before you've even merged it yet and then if you're happy you click merge then the real continuous delivery pipeline kicks in it is a real release with a new version and it goes out to prediction with human approval so that's pretty much all I have to say I guess about upstream so fabricate is the upstream I guess nothing I should mention is what's awesome about fabricate is we're looking at the complete end-to-end development cycle for you offers from analysis planning coding development testing staging operating and so forth the entire end-to-end soap replicate is kind of huge because it's doing kind of everything so it's got a it's got planning in there workouts and tracking in there it's got the eclipse cheer integration an IDE it's got technologies for doing builds and deployment and staging and complete integration and conduce delivery it's got all the analytics that you saw in there for analyzing stacks upon them it's kind of huge but what I would say is it's an awesome open source project it's the most fun I've ever had in open source it's amazing please join and have a play if you know any developers tell them to do it as well it's awesome if you're wondering we're about to start the bit I'd recommend it's a stuff I work on mostly it's a continuous delivery pipeline section right we have an amazing library of pipelines that do again automating build staging system testing integration testing load testing unit testing and whatnot and rollouts what we have is pretty awesome but I'm sure you can think of new things it could do I'm sure there's different things you'd like to do in your pipelines and show there's different tools you'd like us to use so that's the perfect place to start where you can fork all of our libraries hack around with it and submit your public recipes or issues I think that's all I have to say about propagating C all right so we still have a few minutes so I imagine some people may have a question or two so why don't we see there's a question right there do we have a hand mic or anything we're just going to I'll just repeat the question so the question is will we have a product eyes on premise enterprise offering so the plan is yes to do that when is when we feel the product is good enough to do so I mean it's certainly lacking like it doesn't have an integration with like enterprise active enterprise directories LDAP and things like that so there's some things like that we need to do when you when you have the product and you bring it on from yes we assumed it would require to be running on open ships that's where we build it that's how we run it so that's kind of how we support it but that would be the essential model we would have there we currently run this on two OpenShift clusters so we have one OpenShift cluster which it's kind of where all the shared services run so the UI - served out or the planning database runs the tenterhooks the key flow the episode runs in our shared space and then we also use OpenShift online to run all of what we call the tenant services so your che your Jenkins build pods they're all actually running in the the OpenShift online cluster and that's running in your namespace so in fact if you know we didn't show this but if I go here and go here go here I can get to the ownership console for login we've got the SSO work that's good and you can go to the project you can see in fact these are all the different protein shift projects I've got so you can see we're running chain we're running Jenkins we're giving you the different environments like run at stage so you know I mean if you want things to work I wouldn't recommend going in there and fiddling around with these things but you can actually go in and look at the way the deployment context of setup or the pipeline's are set up in here we're giving you access to all of that all of all of your services are kind of running in your openshift online yeah other questions yes yes yes you can configure incurred so the question was can you reconfigure the commit is this genitalia are you tool the stuff will stay around for the next store cause that's gonna be all about you yeah but yes actually you can customize yours chair stacks to be whatever you want put any tools in there like what with any programming language you like what with any compiler tool the one that we give you a lot of rope to hang yourself with I mean it is you know we we didn't talk much down in here because like I said as open-source developers ourselves we don't like it when things are locked down we like it that the happy path is super easy and anyone can navigate it but if we want to get into things we want to get into them and be able to make those kinds of changes so yeah only restriction and the reason why when Todd said can you young instill things and I said no it's that we don't run containers with root privileges we think that's really important so while some inside that container there I can't install anything but the chase stack is a docker file and in the docker file you can get root privileges and so run yarman and install other things that's how we install articles into the stacks right yeah that's the the fabricator analytics plug-in that you saw running in enchain in the keynote that's simply something we install into the container for you other questions yes please what languages do we have Anna will explore what are the plants so we so in open shipped out i/o today the only analytics that are turned on for normal people that go and use it is for Java we already have analytics working for node Python yeah so we have good analytics tonight and we're working on Python at the moment and you know we and we got a road map that hits you know gems and everything else right so we'll kind of knock them down in kind of popularity order and and the intention is like oh well this is version 1 and we'll roll those out in version 2 like in a year we don't view it that way actually the version number is well it's in the about box I get changes daily it changes hourly right so we don't talk about version 2 or 3 we talk about version 7:18 a.m. today right because we we're constantly adding new features in capabilities so this is you know 20 17050 1 that was yesterday right today yesterday my watch is off at 1 to 100 well let's twitch 100 UTC UTC okay people were still working like it really and if you go that about blocks later today it may be changed again so so yeah and but honestly understand this that the most important thing to watch is the community around this so it's really not what we decide should be the right next language or the right next to the panel is when you all decide to the right next language right analytics right you all have an equal ability to be part of making that decision as we do right and that's it's super important to us you know an intra Krishna who leads the team for that part is talking in this room I think it the last session like I think 4:30 p.m. and one of the things that I was going to mention there is that when you get access to the service one of the things you get is an access token and there is a full REST API for the analytic service so you know we're just to build these things we just call that API we don't we're not got some sort of secret lurking or something like of you if you were really ambitious and one too tied into some other tool that you had you could absolutely and and he will be showing some stuff that's coming absolutely so that every talk I think will have the view to the future versus what's just available today yes so the question is how does it work with various doctor registry then can I extract the container image to run somewhere else I'll go for this one please so we're doing all releases and continued integration using a Jenkins pipeline the Jenkins pipeline is built using something called the Cuban eighties plugin in Jenkins which uses a Cuban eighties pod and OpenShift to do the build so the builds are all running in a pod then that build using something called source to image and open ships to make a new docker image it's using something called image streams and openshift to create the new docker image that image stream is then used to download the image onto every host that runs the application so we're using the docker registry of expecting to OpenShift and every image is then versioned and tagged with a new release version so every time you do a continuous delivery release as a unique version number and then unique docket image the tags with that version number in the opportunity stream yeah you can really that download it locally absolutely by just using the open shape docker registry directly in many ways if you're running it locally in CheY where you saw paid using j that's actually still running on openshift even though it looks like you're running it locally so if you're using j you're still running everything in containers on openshift be absolutely can run it locally you also can run it without building the docker image as well if you if you want you could do B then maven vertex run or whatever if you want to run it that way if you're on a laptop and you want to build and run it and using Java and maven you can use maven fabricate Cole and run and that will build it and run it locally but it actually runs it on Cuban 80 so it's not running actually on your local machine it's running remotely one of the other choices are running actually on your laptop which you don't really want to do but you can if you really want or you can run it locally on Cuba native and in terms of the images themselves so one of the things that's on the roadmap is right now it is key if you use the pipelines here they deploy to overshift online that's what it knows how to do the goal is to have it to deploy to OpenShift anywhere right so regardless of where you're running open ship Ian you're running it locally you're running it on another cloud provider you're you're using open ship dedicated you know the idea is to within a short period of time be able to target all those because your environments might be different clusters right it's very common for people to have a different prediction cluster right so development testing and staging might all be on one officer course then you have another what's your cost for production maybe yes please yes yes absolutely so everything you see here we're actually using this to build itself so we're using fabricate we're running fabricate on the Kuban if you cluster and then using that to do all of our bills and deployments and runs of this so absolutely you can you can use fabric upstream yep so yeah so the if you wanted to go do it yourself and contribute in the upstream and be a part of making it run locally it's physically possible but the easy thing to do is go to over super easy all right so we are at time I believe we're maybe one question that would be quick about that how does it integrate with cloud forms in many ways it's more houses clubs phones integrate with each person uh everything runs on openshift and then club forms comes integrated with open ship to more than to manage everything on local ships so in many ways cloud forms takes care of that forwards which is awesome so we'd have to worry about it we really felt we really try to not get rid of and high technology doesn't need so you know get is fully available right so you can pushed up into get we didn't hide open ship so you can still use the open ship cloud forms integration the you know we're trying to just build on top of what's there and make those things available in different UI 3 yeah as an example this build pipeline I have running here if I go into my project here and I go to my pipelines here you can see it's exactly the same pipe tied in exactly the same state you can use either UI to get to it it's just the same data but presented in slightly different ways as we think different people appreciate data in different ways you know I don't necessarily want the things that are here I want other things if I'm a developer sex so I want to be a you know preachers of people's times because if you're going if you want if you're going to stay for eclipses Shay keeping your seat here in a good place but if you want to go to a different session we're five minutes over but so we'll give people a chance to without thanks so much please do

Channel: Red Hat Summit

Views: 10,917

Rating: 5 out of 5

Keywords: Red Hat Summit, OpenShift.io, OpenShift

Id: lkNVTY3vzvs

Length: 45min 30sec (2730 seconds)

Published: Mon May 22 2017