Hacker News new | past | comments | ask | show | jobs | submit login
Red Hat OpenShift.io – Free, end-to-end, cloud-native development (openshift.io)
276 points by artsandsci on May 2, 2017 | hide | past | favorite | 113 comments

I wonder if there's a name for this: that impatience and discomfort you feel about anything new until you understand what it is exactly and what to compare it to.

Until that question is answered, you have ZERO patience for any marketing you read: especially fluffy visionary commentary about "digital transformation". You are quite literally making your readers angry because you are not answering their questions.

For anyone launching anything:

1. Make the first goal of your homepage just explaining what the thing is. Be clear. Be specific. People just want to know how to reason about it.

2. Show photos, video, or screenshots. (Even if you don't think it's amazingly sexy.) Seeing the product makes it so much more real.

3. Compare it liberally with what's out in the market now: this is your chance to show how it's similar and why it's different. Can be chart or table.

Then and only then, can you move on to how your product improves the lives of devs, chefs, athletes, etc. who uses it. (But you may not even need to.)

You seem like one of those guys who simply can't appreciate a vertically integrated, synergistic solution that delivers on the promise of strategic technical innovation with seamless delivery.


This is my biggest gripe as well, and you're spot on. Vague promises of how the thing is going to "revolutionize the way I do things" just make me file it under "things to impress enterprise middle managers with", rather than something I can use to make my life easier.

It's a hard problem, especially for a product that has as much surface area as OpenShift.

- It is a PaaS, which would compete with Docker, EC2, and Heroku.

- It's development environment. Sublime?

- It's a collaboration tool. Github?

More to the point, it probably doesn't do any of those things in particular quite as well as those specific solutions do, but they're hoping you see the value in everything together. Would there be value in an IDE that had a deploy button that automatically marked a task as done?

I'm not sure that's exactly it, and I agree that this site isn't wonderfully done, but it's hard to get across "the thing" when there isn't a perfect comp to "the thing" already out there. Bad comparisons will pigeonhole your product forever.

At Pivotal we call this idea the "Circle of Code"[0], which lends itself nicely to a diagram.

We see similar value in it, but so far I've seen it explained more via products that play nicely together versus packaging a single platform.

I guess the question to ask is: if I take away this part of the circle, how much worse off am I? The anticipated pain gives some inkling of how important it is to provide it. Take any of Tracker, Cloud Foundry or Concourse away and it all becomes much harder.

Where I personally differ is that I like plain ol' IDEs and text editors. The web-based IDE thing keeps coming up. It may be that it's one of those ideas waiting for its time. Or may be one of those ideas that is a bad idea. I personally think the latter, but my predictions about technology are so frequently wrong that I should write for Wired. (Edit: it's also possible that I don't understand what Eclipse Che is)

Kudos to Red Hat for tying it all together. People forget how many smart PMs, designers and engineers love working there.

Disclosure: I work on Cloud Foundry on behalf of Pivotal. I'm just an engineer, I shouldn't be cited as authority about plans or whathaveyou.

[0] https://youtu.be/7APZD0me1nU?t=23m10s

Hi @toddmorey - I am founder / CEO of Codenvy. Eclipse Che is our open source kernel and we had the opportunity to collaborate intimately with Red Hat's engineering teams in the design and development of OpenShift.io.

Let me share some of the perspective that we have on what they are building.

What happens if your app dev lifecycle tools (git, issue management, IDE, CI / pipeline) runs on the same platform that is running your production applications?

From a market point of view, openshift.io is an approach to developing containerized applications in a way where the development team does not need to setup the infrastructure and tools required to create those workloads.

Red Hat has:

1. Created their own agile issue management offering, competitive to Jira, GitHub issues, or Gitlab issues, based around the concept of a "work item".

2. A context-aware browser IDE based upon Eclipse Che, which "dev modes" a workspace based upon the containerized version of the application you are building. This injects things devs need for writing code including SSH, language intelligence like auto complete, debuggers. The workspaces are configured for the developer in-context based upon the work item that is under taken (ie, which branch, which repository, which languages, which files are referenced).

3. A continuous integration pipeline designed around building, testing, and deploying containerized servies built for the project using pipelines tailored for the team and for the individual. The pipeline capability is built from within, but takes advantage of a containerized Jenkins under the cover.

4. Deployment of the application to OpenShift. The application is packaged as a set of containers, which (OpenShift) is powering all the underlying services (Che workspaces, work items, pipelines, etc). The deployment automates complexity about how to set up a deployment profile, automatically adds canary deployment mechanisms (deploy for only the developer, for the team, or for production users (portion of users, etc)). Why not use canary deployment for managing the builds of commits for individual engineers along with your end users?

5. Analytics built in for doing intelligence of the code base that is being built to do analysis of the open source libraries included in the project you are creating to identify if any of those libraries are potential security issues, as identified through community flagging or a white / black list provided by an enterprise.

6. By building all of this on a common underlying infrastructure, there is additional context associations that are understood, so pipelines are connected to work items and the workspace, and vice versa. This alleviates certain overhead and setup requirements of developers long term.

I am excited personally about this product because it solves the biggest pain that Codenvy sees with our paying customers. We have a number of F500 customers who pay and deploy a private Codenvy.

But they are all becoming DevOps infrastructure wizards. It's hard for them having to figure out how to wire together Jenkins, Gitlab, Kubernetes, their IDE. They can do it, but they are seeking a DevOps in a box solution that makes that sort of setup effortless for the dev team, self-service (so that the DevOps team doesn't have to spend weeks getting it configured), and production grade (lets dev teams build enterprise applications that will be packaged as container workloads).

So openshift.io is what happens when an organization tries to remove the effort out of DevOps, and they ultimately make optimizations and workflows based upon that point of view.

I am also project leader for Eclipse Che, and we are making a commitment to migrating the management of our open source project (4K github stars and >100 contributors now from a diverse set of companies) onto openshift.io. It will be really interesting.

Whoever is in charge of openshift.io really ought to just delete their landing page and paste this post in its place. This is a thousand times more helpful than the meaningless buzzword soup that's there right now.

Thanks for the detailed writeup. I totally agree—there are corners of the product that seem pretty exciting, and I'm certainly a fan of redhat. I just think they could really help the effort by explaining the product better in their launch materials.

One quick question for them to address is if the entire stack (suite?) will be runnable on your own hardware or if it can only be consumed as a service.

I have bumped into all of the OpenShift.io managers and executives today at Red Hat Summit. They have all been discussing this very HN thread (they have all read it), and taking it to heart that they want to clean up the marketing items immediately.

As for your question - initially it will be released as a service. But it's Red Hat, and their customers are primarily on-premises with OpenShift sales, so the demand for on-premises deployment on your own hardware will be through the roof. So we'll see if RHT targets that demand.

We will. (12 year Red Hat veteran speaking)

I have had a chance to write a proper blog post about my views on OpenShift.io and what it does, and how it makes use of Eclipse Che.


This is really neat.

And the "Learn more" link for every single tile and bullet point on the home page. How can someone in their right mind think "Lets scatter the information on the home page into 20 links so the user can click on them one by one and handle the navigation themselves"

It's what happens when you write your content for search engines instead of humans.

> 2. Show photos, video, or screenshots.

You'll literally be able to demo it yourself for free, on demand, in production (IF YOU DARE!), within the next few days.

Well, that's still not the same. Pictures, videos, etc are supposed to give you an impression what the product is, how it tries to solve that problem, etc. Before you invest anything in it. Even the time investment of singing up, and playing around in the "member's area" can be a hassle if you could tell that this project is not for you from a simple screenshot.

Furthermore, I would probably have to read the manual, etc before I can do anything useful with it. A good screenshot or even video can really help you with that impression

Then they should have announced it when it was ready.

There is now openshift.com (enterprise), openshift.org (OSS), and openshift.io (???) and none of these websites do a great job of explaining why you should be on that website and not the others.

You want everything explained, you want to use it for free and you don't want to participate. Everything that's wrong with the world ;-)

Don't be silly.

If you want customers, you have to actually convince them to try your product.

The expectation of a customer that they might want to understand what the heck they're signing up for... before they sign up for it is hardly controversial.

I think this thread makes it abundantly clear that a days work creating some screen casts instead of fluffing out some marketing fairyfloss would have significantly improved the reception of this product.

It's just plain marketing fail.

Why would I be digging through Red Hat's blog to find this, and not find it on the site of the actual thing itself?

I was expecting to find a video in the link you pasted, almost missed it. Here's the actual link to the video https://vimeo.com/215402513/02867b4aea

That video adresses a bunch of the issues noted. It briefly shows provisioning, the editor, deployment, issue tracking in action, and the narrator overviews the vision & reasoning. It should be touted more prominently.

Then wait a few days and provide a video-backed live walkthrough.

The intro is intended as a teaser for the demo. If you don't show people what it is, they won't demo it.

A missed opportunity, but there is still time to recover.

Seems like people haven't used openshift. It is fun, it is easy, and it is free stuff. Pretty much a replacement of aws do azure gc etc. Atleast for poc and mvp. And if u r smart enough then u can process millions of request per second for free in ur choosen stack.

What is this product about? I clicked "Learn more" and was greeted by a sweet, juicy corporate-speak (https://www.openshift.io/end-to-end.html):

    Digital transformation is about evolving into a technology
    business in order to compete in the digital economy.
    Businesses can’t transform without relying on the developer
    to implement the transformation strategy and deliver value.
Mmm... tasty!

I think with some stake-holdering it can synergize the democratization of orchestrated micro-services cloud.

That's old school, you didn't transform a single thing.

but is it web scale?

In real-time.

I wish I could join in, but this stuff is par for the course. Every time I raise A Platform I Am Notorious For Mentioning, it is pointed out to me that the website is kinda indirect and that my employers' website has a similar degree of obliquity.

I expect Red Hat engineers will be here to help us parse it.

Disclosure: I work for Pivotal, we compete, etc etc

I am glad im not the only one confused by this. I went to the webpage first, couldnt really figure out what exactly it did. So i went to HN comments to see if i could figure out the purpose of this product through the comments. Turns out everyone is confused as well, lol.


OpenShift is a great PaaS that brought a lot of innovation and improvements to Kubernetes, congrats team on the awesome and useful product!

Here are just some of the influences of OpenShift on Kubernetes that I'm aware of:

* RBAC model (https://github.com/kubernetes/features/issues/2#issuecomment...)

* Kubernetes notion of namespaces was largely influenced by RedHat and Open Shift team experiences in the enterprise AFAIK

* Continuous ongoing work by RedHat team on both OpenShift and Kubernetes on the security and performance (https://github.com/kubernetes/kubernetes/issues/12742)

I'm sure there are many more improvements to Kubernetes to OpenShift that I'm not aware of.

So... it's like a poorly explained PaaS which is:

    - An IDE  
    - A platform for running containers
    - A set of productivity and collaboration tools
    - CI service
    - Deployment tools
...all in one?

Really? I'm scratching my head over this one.

I really respect the work RedHat do, but there's this idea of picking one thing and doing it well... but heck, I mean, you can also just do everything as well, and hope one bit or another sticks I guess.

? ...but why would you? How can you expect to offer all of that at any kind of acceptable level of quality?

Why would I pick this up? It feels like if you do, you're basically going to play in some walled garden where the rules might change at any time, and you'll be out in the cold when they do.

Full Disclosure: I'm a Red Hat Consultant focused on OpenShift.

All of our work is open on github under the OpenShift Origin project, Kubernetes, and other projects. We push most of changes upstream.

Doing containers at a scale of 10,000+ running containers would be unimaginably hard without something like OpenShift. Clusters of 1,000s of nodes, Security, RBAC throughout our supported tools, optional SCM (git via GOGS), optional integrated local Docker Registry, optional CI/CD (Jenkins), optional build/artifact repo (Nexus), optional basic metrics (Hawkular, Heapster, Cassandra), optional monitoring (CloudForms, Cockpit), optional EFK logging stack, (Elasticsearch, FluentD, Kibana), horizontal autoscaling, support for NFS, NetApp, OpenStack, RHEV, KVM, VMware, AWS, GCE, Azure, Ceph, GlusterFS, BIG-IP F5, iSCSI, FibreChannel, dynamic node provisioning, dynamic storage provisioning, OAuth2 authentication and app integration, LDAP authentication, SAML2 authentication, Okta, auto SDN creation, SDN segregation, application load balancing, multiple encryption flow options, HTTP/2 support, internally secured communications, Ansible integration. For a solution so encompassing, you need something that marks the way. Most of our customers exchange several parts for what they have or what they want.

And we do all of this with Quality as we develop in the open & commit to upstream projects. We spend $100Ks (probably more) on R&D and testing every release (every 3-4 months). The benefits, again, are usually committed upstream. You have Red Hat's award winning Customer Support. I'll personally vouch for our OpenShift Customer Support Teams. Absolutely amazingly talented and insightful engineers from all over the globe.

One of our biggest customer concerns is in integrating customers' code to run on the Platform, aka, the Last Mile of Integration. IMO, OpenShift.io is intended to reduce that barrier. We support just about all the popular languages, frameworks, and platforms. We have examples & templates for Perl, PHP, Python, Ruby, nodejs, Java (Spring Boot, JBoss, xPaaS, FUSE), and even .Net Core, MySQL, PostgreSQL, MongoDB, CouchDB. Figuring out how to integrate legacy CI/CD, monitoring, logging, IDEs, etc. with OpenShift takes the most time. In choosing OpenShift.io, in return, you get an environment that is built from the get-go to work together.

Almost all the internals are matured/stable upstream projects which have been around for a while.

Openshift.IO makes all them work as a SaaS for production use.

Last I checked, OpenShift was comparable to Heroku.

Not quite. OpenShift 3 is oriented around containers as the unit of deployment and operations. Heroku pioneered the buildpacks model.

I believe OpenShift has a buildpack-like onramp now, I am not qualified to name it or describe it.

Deis takes the Heroku buildpacks from upstream and so is more-or-less plug compatible with Heroku. Cloud Foundry does something similar, though with extra wrapping to allow it to work smoothly in disconnected environments. Some of those buildpacks are being rewritten, but I expect they'd be compatible with Deis too.

Disclosure: I work on Cloud Foundry for Pivotal. I've been on CF Buildpacks twice.

Yes and no. Deis Workflow is like an open-source Heroku (even uses Heroku buildpacks.) OpenShift has its own build stacks and does not share as much in common with Heroku and Workflow.

From a naïve user experience perspective, OpenShift is probably a bit more like Heroku. Workflow does not (yet) have service brokers or add-ons that provide sidecars like a MySQL database or a Redis server. OpenShift provides those things in a manner more similar to Heroku's add-ons.

OpenShift also provides users with a point-and-click interface, but Deis Workflow expects you to use a command-line interface or bring your own dashboard (deisdash is a thing).

Deis Workflow depends on Kubernetes, and OpenShift =~(subsumes/supplants) Kubernetes in a marginally incompatible way, but to my knowledge Heroku has no relationship at all with Kubernetes. At least if it does, it's very opaque about it. (I couldn't tell you either way because it is opaque.)

Can someone explain what's new here? Openshift has existed for a while, right?

Edit: Thanks. Feels like better wording on the page would have helped. "Online IDE" or something. Or rolling screenshots that show what it does, like Cloud9 has.

Or maybe I'm too old to grok all this new cloud terminology.

disclaimer: I work on openshift.io, and there will be much more coming out as summit goes on so I can't refer to more than in current news:

From https://www.redhat.com/en/about/press-releases/red-hat-unvei...

  Team Collaboration
  Agile Planning
  Developer Workspace Management
  Application Coding and Testing
  Runtime Stack Analysis
  Continuous Integration and Delivery
thus it is more than "just" an online IDE in the traditional sense of an IDE, i.e. Eclipse Che is just one part of it.

All of it running on and targeting OpenShift.

Hope that helps.

Sure. I wasn't trying to minimize it by saying "online IDE". I was trying to say that the home page is pretty ambiguous. I needed help from HN'ers to tell what it even was. "OpenShift" already means something else to me. That list of features you just posted would be good to put on the home page.

I've read through most of it, and I think you're safe calling it a web-based IDE. Obviously "IDE" means more than it would have 15 years ago, but that's still what it is.

So if a company chooses openshift.io, then I'm not going to be able to use IntelliJ, NetBeans, emacs, vim, visual studio code, etc.

That'll help attract talent.

Proof for your statement? This sounds like juicy, paranoia-mongering FUD.

Sorry, I didn't mean it to come off as a statement. It says that there's an online IDE, so it was a question; I unfortunately used a '.' to end the first sentence because I had an 'etc.' at the end.

It's seriously a question; and if the answer is yes, then the following statement applies (the one about attracting talent).

But yes, it came off pretty cynical and sarcastic, but if someone is going to try to force me to use a particular IDE, it would piss me off (and the thought of it did too).

If it works anything like OpenShift 3's on premise product, and I see nothing that says it doesn't, you can use whatever IDE you want, but the deeper you get into customization, the less "special sauce" you'll get. Maybe clippy won't pop up and say "hey, it looks like you're trying to use an out of date library!" if you're editing in vi and checking in with git or whatever.

I think you are jumping to conclusions here, OpenShift.io is committing to a git repository. I'm guessing you'll still be able to clone this repo locally and use your own tools.

Yeah, I don't see any reason why you'd be forced to use any particular IDE if you're using Git properly

You can use any IDE and also use every other component of OpenShift.IO ( like planner, build, pipelines )

I have an OpenShift account and now I'm shown an "additional action needed" page on login. When i complete it and hit submit, it tells me that email already exists.

Not at all pleasant :(

I'll see if we can get a fix pushed out for this today/ASAP. The issue currently occurs when you have multiple Red Hat accounts that use the same email address (or a sub-addressed version of the same email address, like "chris@gmail.com" and "chris+123@gmail.com".

> I'll see if we can get a fix pushed out for this today/ASAP. The issue currently occurs when you have multiple Red Hat accounts that use the same email address (or a sub-addressed version of the same email address, like "chris@gmail.com" and "chris+123@gmail.com".

Is this a special case for Google mail or do you assume foo+bar@example.domain is the same as foo@example.domain ? Why?

It's an abuse prevention mechanism. We do not permit users to have multiple OpenShift Online accounts. See 2.1 in the OpenShift Online terms of service: https://www.openshift.com/legal/terms.html

"You may not (or permit third parties to) create multiple accounts or otherwise access the Services in a manner that is intended to avoid Fees or to circumvent maximum capacity thresholds for the Services."

Then it's a bug right? The uniqueness check should be based on the part before the + sign.

No, I don't think what google does is standard behavior (not trying to say anything bad about Google; I love Google). However, as far as I know, disallowing + is not a particularly good way of preventing abuse. I won't go as far as to say it semms like bad code but it is pretty close.

I think it is OK to make a special case for Gmail and Google mail but it is counterproductive to try to crown it as a de facto standard.


huh? Google's special behaviour is that they are blind to dots in the user part of the address.

The + sign is not anything to do with Google.

Oh I didn't know. Just found https://tools.ietf.org/html/rfc5233 after trading what you wrote.

I just assumed + is a valid character and a+b@example.domain would be a valid email address independent of a or b. But it seems we shouldn't allow + or -- when people sign up for a new email address?

Sorry for the inconvenience! I have reported this issue here: https://github.com/openshiftio/openshift.io/issues/134

This issue has been fixed in upstream. https://issues.jboss.org/browse/RHDENG-1320 Please try it again after some time.

We've pushed a fix for this. Please give it another try.

This looks like a clusterf*ck.

The upstream for openshift.io is primarily in the fabric8 project. fabric8 is found in the following organizations:

https://github.com/fabric8io https://github.com/fabric8-ui https://github.com/fabric8-analytics https://github.com/fabric8-quickstarts https://github.com/fabric8io-images

OpenShift has existed for a while. There are basically two versions; the original version was like a Heroku competitor. Then about 2 or so years ago, they came out with a new version based on Docker and Kubernetes called OpenShift Origin (OpenShift 3). I bet this is some type of OpenShift 3 offering, but it would be nice to know how it differs from their current offerings.

Disclosure: I work on openshift.io

Openshift.io is the end-to-end development environment which is built over the OpenShift PaaS.

Sorry, but you are doing a terrible job of explaining what OpenShift is to outsiders.

OpenShift is a Platform-as-a-Service offering from RedHat. It provides a fully integrated solution to building and deploying applications on Docker and Kubernetes. What makes it a PaaS is that it integrates Jenkins, so you can build images from source and promote them though environments. https://www.openshift.com/

OpenShift is a private cloud offering, so it is something you install and run yourself on your own infrastructure or in a public cloud.

So OpenShift.io is the "battery-included" version that is already installed on a public cloud, managed by RedHat?

> What makes it a PaaS is that it integrates Jenkins

Not just Jenkins. There is a Jenkins button and you can use Jenkins for CI/CD, but you can define BuildConfig and ImageStream through their point-and-click interface without ever adding Jenkins or writing a line of config by hand.

You just need to use one of their built-in ImageStreams and DeploymentConfigs. I got my first exposure to OpenShift through the Developer Preview/OpenShift Console beta. At the time I was testing it out, deployment of Jenkins was outright broken, but I was able to get my app working and builds automatically generated in response to GitHub webhooks (eg, CD without Jenkins.)

The Developer Preview was a very interesting experience, and their support people were quite responsive in #openshift-dev, but I'm giving them some latitude in that it was clearly labeled as Preview / Beta and a totally free product with built-in expiration date.

I would not have been likely to call it a good experience except that it was clearly labeled beta. But if it says Beta on the tin, and I am able to complain until it works, that's about where I set the bar for good experience. It was a good beta experience.

If I had to guess, OpenShift.io is the next phase of this:


(I am certain that you will not need to use OpenShift.io online IDE to get "batteries included" OpenShift from RedHat.)

Ok so this is "OpenShift.IO" :) And not just openshift

This is openshift + planner + code editor + pipelines + 'analytics'.

Here analytics does code quality checks and lets you, the developer know, if they should be using a different Jar/package ( example: the one used by the developer might have a security vulnerability reported which the developer was unaware of )

They should have paid you to write copy for their website.

This looks to be very much like codenvy.io or cloud9. The change seems to be that it's built into the stuff(CI/CD/containers/*) openshift provides.

Looks like it's built on Eclipse Che - what Codenvy is also built on.

Folks, check out the TechCrunch article on the announcement :


It will answer a lot of questions :)

Can't even sign up, not a great first impression for something that I will have to answer to my customers when it doesn't work, but it's free so I have no right to complain right?

Service Unavailable - Zero size object

The server is temporarily unable to service your request. Please try again later. Reference #15.4789fea5.1493742308.9b867eb

Internal Server Error - Read

The server encountered an internal error or misconfiguration and was unable to complete your request. Reference #3.5689fea5.1493742329.8e1cbf2

I signed up no problem.

the dumbest authentication process I've ever seen. I can't remember my password. Click forgot password. Click on email. Enter screen to change my password. Submit and it tells me it needs more information. Type more information. Complains the email already exists. Obviously, I'm resetting it. PRetty stupid.

We've pushed a fix for the duplicate email issue. Please give it another try.

I just hit this too. I guess I have to abandon that account and use a different email.

openshift.io (please watch .io) is a complete cloud native development environment for organizations, having work spaces created for collobaration, IDE with Eclipse Che, complete build, pipelines along with analytics to deliver actionable insights to the developer for choosing the right software components. All this happens in a Browser..

I run the group at Red Hat responsible for OpenShift.io and our developer program. We really value the feedback and the comments on this thread. I admit, we could have done a much better job explaining the product on the site.

Based on everyone’s feedback, we redid the homepage last night. We hope this does a better job explaining the benefits and why you would want to try OpenShift.io. Please give it a look and let us know how we did.

I know there are a lot of people who want to give OSIO a try. We will be onboarding customers as soon as we can. We are putting in the finishing touches and ramping our capacity to meet the demand. We will be keeping everyone updated on our progress through email, twitter (@rhdevelopers) and our developer blog - https://developers.redhat.com/blog

I remember I was really excited about cloud9 online IDE, but it took me about an hour to get back to usual development. So let's give this one a try and go back to usual ones :)

Here's a video demo of OpenShift.io by Red Hat:


Has anyone signed up for the preview? I keep getting "We're sorry, but something went wrong." right after logging in.

We're getting a fix out momentarily - it's a character encoding issue in our DB (default MySQL UTF8).

In the meantime, replacing characters in your name like "Ł" with "L" will fix the issue. You can update your name and account settings here: https://developers.redhat.com/auth/realms/rhd/account/

The auth process is stupid. Can't login because I don't remember my password. Can't reset because it needs more information. Trying to hook up with Github doesn't help either. Very frustrating.

I did, but now it's in the status 'waiting for approval.'

Still waiting for approval.

I am as well too still. @steevenwee any word on how long it typically takes?

I'm stuck in and endless loop of "We need you to provide some additional information in order to continue." and "Email already exists. Log In".

Hey! Sorry about that, we are working on fixing the problem for folks who had an existing OpenShift account. Tracked it here: https://github.com/openshiftio/openshift.io/issues/134

Same here, already had an openshift account.

Signing up with Github went flawlessly for me.

I can't figure out what this actually is, past all the marketing speak. Picking the first box on the homepage ("cloud-native development experience") gave me no actual information about that experience entails. Yes, yes, it will unshackle my free time and relieve me of madness, but how?

What is it?

C'mon, RedHat. Throw me a bone here.

MVP edition for web site testing posted today by christux:

Show HN: Tux.io – a (now working) Linux desktop in your browser | https://news.ycombinator.com/item?id=14245447

How does this work with rkt?

rkt is still not fully functional as a container runtime for kubelet[1]. kubelet is the underlying daemon used by Kubernetes and OpenShift for running containers. I am not sure if and when rkt will reach feature parity with the docker runtime but once it does, we wouldn't need to do much and enable it in OpenShift.

[1] https://kubernetes.io/docs/getting-started-guides/rkt/notes/

Why do I need to provide my e-mail and create a password when I choose to sign in with GitHub? First of all they were supposed to be able to retrieve the e-mail address from GitHub I think, like they did with name and company. Creating a password should not have been needed at all.

Tried OpenShift last time and their cartridge model didn't make sense. What has changed since then?

Literally everything. Cartridges are OpenShift v2, "images" in v3 are literally the same concept as a Docker image. You can check out this link[1] for more about what's changed.

My first instinct trying OpenShift v3 Developer Preview was to slap helm on it and see if I could deploy my Kubernetes projects. Turns out that Helm does not have much of a security model to speak of, and this was problematic to enabling support for Helm in OpenShift v3 where multi-tenancy was one of the concerns they wanted to address right away.

I got with claytonc in the #openshift-dev channel on freenode and he explained all of this to me, the long story short is that Helm on OpenShift is not well supported but that you can use it if you are willing to give it a service account with editor access to the whole cluster and make some other small changes like setting TILLER_NAMESPACE to something other than kube-system.

Instead, OpenShift basically uses the Kubernetes model with some additional primitives for deployments and builds. You define a BuildConfig (or let the point-and-click interface do it for you). Builds produce images, and images go into an ImageStream. A DeploymentConfig maps deployments to ImageStreams. [2]

The rest of the way down the stack, your deployment looks very similar to a regular Kubernetes deployment (except it's not using the Kubernetes deployment primitive, I believe that OpenShift implemented their own formal concept of Rolling Updates on DeploymentConfig before Kubernetes released the modern Deployment/ReplicaSet primitives that many k8s users are using for this now.) It is almost Kubernetes under the hood (but not quite, in subtly incompatible ways that depending on your size and current level of investment into Kubernetes, you might have an easy time or a hard time getting over.)

If you're looking for a short list of things that have changed from OpenShift v2 to v3, there isn't one because they are not really directly related products other than by name[3].

[1]: https://access.redhat.com/documentation/en-us/openshift_ente...

[2]: https://docs.openshift.org/latest/dev_guide/deployments/how_...

[3]: http://stackoverflow.com/questions/32477650/openshift-v3-vs-...

TL;DR in a software defined everything world, Open wins. Use it. Enhance it. Participate. Disclaimer: Red Hat guy since 12 years.

It's Redhat, whose focus has always been enterprise the money maker.

Ever since it dumped the desktop years ago, I switched to the deb ecosystem and never looked back, do not want to have anything to do with Redhat,including this openshift thing, one reason is for the "dump-the-desktop"(please don't mention Fedora), another is that, DEB is so much better than RPM for my daily life.

I also hate the Redhat-intrusively-forced-on-SystemD to my heart.

That been said, I appreciate what Redhat have been doing to the open source, I guess it is win-win for the community and its own business goals.

By the way, my 2017 wildest forecast: Microsoft may acquire Redhat

Red Hat is to some extent protected from acquisition by the ideological poison pill. Because there's no IP (open source), the value of the company is basically the engineering talent, which enables the level of service Red Hat provides. Many of those engineers have very strong opinions about FOSS.

The second a Microsoft (or worse, an Oracle) tried to acquire Red Hat, a huge portion of that talent would walk out the door.

No, the value of RH is in the clients they have, just like every enterprise software company ever.

Oracle is notorious for buying companies for their clients, even if everyone quits (eg, hostile takeover of Sieble)

There is more chance of the pope converting to Islam, then MS acquiring Red Hat. That is one acquisition that would turn to shit over night, and MS would be left with just some logos, as the companies main assets (its engineering) would all walk out the door, followed by customers not renewing subscriptions.

The Microsoft and Red Hat subscriber bases would have massive overlap.

It's 2017. Pope converted to Islam and Trump got elected. Didn't you hear?

2017 forecast... Microsoft acquires Canonical.

that will be the worst, please don't, of course we still have debian.

I've moved to Arch. They can have Debian too.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact