Red Hat OpenShift OpenShift 3 Demos, 7 Jul 2015

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: OpenShift 3 Demo Part 8: Rolling Deployments

Description

In this video, Veer Muchandi demos rolling deployments on OpenShift 3.

NOTE: For the latest information on OpenShift 3, please visit https://enterprise.openshift.com or subscribe to the OpenShift Blog (https://blog.openshift.com).

A

In this video, we are going to talk about rolling deployments with open shift at this point.

A

I have an application that has a PHP front end and a back end database with my psycho, and this application has been configured in such a way that the source code is is in a github repository and I have taken the web hook from this application running on openshift, and this web hook has been configured inside my git repository so that whenever there are, if there is a commit made to this repository to this project on github, then it will automatically trigger a new build on openshift and also at this point we only have a single part running the front end on an open shift.

A

So when we make any changes to this application, this is the part that gets updated. Now, let us make a small change to this application. I am going to open the DB test, dot, PHP file and add a small line.

A

I'll write and commit these changes.

A

I committed it into the local, get repository and I'm going to push it now.

A

Now, as the changes are getting pushed, you will see that it will trigger a new build, and this is the new bill that got triggered, and once this build is complete.

A

The new image that is created with these changes will be deployed on to the pod that is currently running. So when the yeah it's about to deploy. Now, let me keep refreshing the screen to see what happens to this part. Now the application is getting deployed and.

A

You will notice that there is a slight downtime here and now it's back again, so the application goes down because there is a single power running and that at that particular point of time, when that pod goes down and then and the new coach in the new image is deployed, you see a 503 first for a little while so for this exact reason of that little bit of downtime right openshift provides ab + option for rolling deployments.

A

Now, if you see the number of parts that are running for this front, end is just one if you have a scale application, and you have more than one pod running for this application.

A

When you make code changes and do the same thing that we did before, which is Mako changes and push them, and when the build is done and when it is trying to get lip deployed, it will not bring down all the parts. At the same time, it will do us one at a time or some at a time. So that way it will ensure that your application doesn't go down and you don't get that 503. So, let's look at that now: first I'm going to resize the application to scale to four pots instead of one.

A

So let me do that now.

A

If I check the number of parts there are 1, 2, 3 and 4 and 2 are uh no 2 and 2 or node 3. Now, let's go back and make another code change and push it.

A

Now this time I will go back again and check and the build is now running again.

A

The deployment is now in progress now, let's see what happens.

A

Keep watching the screen, and you will see that this pods are now one of the parts is at least updated. The other part is not updated, so you are seeing this end of the list appearing in the output on some of the parts and not on some of the parts. So the application is not coming down. Instead, the pods are getting updated one after another.

A

You will notice that the end of the list was missing on some of the parts and now now it's like all the pods are now updated and running. So everything is up and running, so I should get the same output, so everybody's updated.

A

So this is the process of rolling deployments where the application doesn't go down completely it does it give a 503 like before when there is when there are multiple parts, the pods are going through a rolling deployment process.