Red Hat OpenShift OpenShift Coffee Break | Red Hat Livestreaming, 22 Jun 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: OpenShift Coffee Break: DBaaS Series - MongoDB

Description

Get your espresso ready for the OpenShift Coffee Break for another episode of the Database-as-a-Service (DBaaS) on Kubernetes at ease series with Red Hat OpenShift! Today we welcome Johannes Brännström, Solution Architect at MongoDB, to show how create cloud-native applications with MongoDB Atlas and Red Hat OpenShift Database Access (RHODA).

Twitch: https://red.ht/twitch

A

A

Good morning, good morning, welcome back everyone to the openshift tv coffee prank, my damn. My name is natalie vince product marketing manager for openshift. Today, I'm with my awesome co-host, hello, jafar.

B

Hello: everyone thanks for joining us again, hello.

A

Andrew hello, everybody thank you for joining us today, uh jafar andrea today we have a very special guest. I'm pleased to introduce you, johannes and johannes, is the pronunciation good bronze storm. No plea, please tell us the right the right one.

B

I I I'm sure it would be easier to set up than than getting the pronunciation correct.

A

Welcome welcome. How are you.

C

Branstrom yeah.

A

Maybe there's a little bit of delay uh but johannes welcome to the to the show today. The topic is um how we get it's part of our database as a service uh services series. uh Today we're gonna talk about how to consume and use mongodb uh from a kubernetes class from an openshift cluster in this case, basically we're going to talk about the mongodb atlas operator and um we will see how to connect the database as a service of atlas. So I'm really excited to for this show today.

D

And I'm, hopefully I will learn something today, because I have never used mongodb. So it's time for me to understand a little bit more about the use cases as well. I can ask some questions.

C

Yeah, I definitely I'm looking forward to that, and also I'm not sure you are uh I'm hearing to hear a lot of echo, uh but I'm just gonna um get started so mostly well. I guess everyone knows openshift and most of you have heard about mongodb.

C

So if there are, let's say three flavors uh of operators that be that we we do with mongodb and really the whole purpose is to to make it easy to use mongodb- and you know I actually used to work for edit as well, and one of my uh kind of ideas of kind of connecting with mongodb is that it really is a database that aligns really well with the containers and an openshift, so um yeah, but obviously uh operation is a great way of running things in kubernetes and openshift as well.

C

So that's why we built them. I'm just gonna. You know uh start with a one quick slide here, so we actually offer three different uh comments: operators, so one to support the community instances. I want to support enterprise instances, which also comes with a tool called ops manager which allows you to do a point in time back restores for backups and also.

C

Database level, encryption and you know really good monitoring- that's bundled uh directly there, but today we're really going to talk about the atlas operator and actually more specifically, the the red hat mongodb atlas operator.

C

So and the reason for that is, we have our mongodb atlas operator. Just went ga like wanted version 1.0 last week and then and the one that's installed in the sandbox which we're going to use today is the version 0.2. So.

C

All right, so this is the openshift sandbox that I've er it actually it's open for anyone to to sign up and use which is really uh really neat and uh as this one loads we'll see that uh the yeah, oh, I think I had removed that. Let me just check.

A

Yeah, this give give us the opportunity, johannes, to talk about what we're going to do today. Now we're going to demo how to consume the mongodb atlas cluster from an open shift cluster. In this case, this is a free tire, uh openshift cluster, that everyone can use. It's called the developer sandbox for openshift. We put the link in the chat.

A

If you like to register it's free to use uh renew 30 days, renewal renewable for free, so you can start deploying your kubernetes apps, with openshift sandbox and in the demo we will see how it's easy to connect databases and database as a services uh using a very cool technology. Jafar andrea, called uh the service binding service kubernetes service binding. Did you I'm curious, jeff arender? Did you ever use the service binding? What do you think about.

B

Yeah, so I used it, I haven't used it very recently, but I liked that idea of making it easier to connect your applications to some back ends just by you know, drag and dropping stuff. uh What so, I think it's really cool, because in terms of user experience, it makes it much easier to consume.

B

The only thing you need to figure out is like how you you set up what environment variables are going to be injected uh when you do that operation, and so I think it's based on some annotations that are some annotations- that you need to put on some parameters, so they get injected etc. But you know that's just the little thing that you have to figure out when you are first getting started with the operator and need to understand how it works.

B

But once this is figured out, I think it's really cool to be able to just you know, uh drag and drop uh these the arrows and have everything set up. Yeah.

C

So am I gonna get started and uh show you a little bit of that um there's a bit of echoes, I'm not really sure.

A

I think did you have, do you have the youtube page also open.

C

No, I only have this page right here.

A

um Okay, that's weird.

B

Especially since you are using your headphones so.

A

um Yeah, that's weird jonas, I don't know. If um did the echo started when we started streaming.

C

And yeah it could start when we start streaming.

D

A

Okay, let's see maybe.

C

Should I try and log back in or okay.

A

Yeah, please try try again to login. um If you we will in the while. We wait for you here and uh uh jafar and andrea um service band, but also it's cool that we can connect the databases as a service database ourselves.

A

So today we're going to talk about two topics: one the in-cluster database that you can consume with an operator, so mongodb or other vendor put the logic in this operator to have your database as a service in cluster and also you can connect to external cluster in an easy way, as you mentioned, jafar with the service binding. So you are connecting this, so you you don't have to write the kubernetes secret. You don't have to write the configuration it's done under the hood and we will see in the demo how okay back.

C

Yes, I'm back and now now it works perfectly okay! So sorry about that, no.

B

Problem, that's better right! So.

D

C

Exactly oh, this is always something but anyway so yeah. We are now hacking kind of a clean slate here and I want to start by uh connecting um connecting the operator- let's say configuring operator, so we're going to use that the provider account and offered by red hat and I'm going to create that and then I need to select a database and provider in this case. Gonna pick the atlas one and there's a set of configuration items there. I'm actually going to just get it from the side.

C

Since it's a credential, I'm not gonna show you that, but.

A

So, um johannes, just to explain we're connecting to the atlas database as a service, is there any free tire that the user can use to run stuff.

C

Yeah there is, and that is actually what we're going to use by the default by default. So this operator is uh by default, creating three tier clusters and uh those are free forever. So it's a great way to to work and play with this, and it's also a great way to to set up your um if you, if you your own projects, for example, if you want to investigate this uh at home, for example um yeah, you can definitely use this for that purpose as well. So.

A

As a developer, yeah yeah.

B

Sorry I was going to ask: how do we subscribe to the account like the provider access how? How do we get that.

C

Yeah, so let's see um what you, what you, what I just did there was: I went in there and put in some configuration details and if I would, if I would do that again, you you see that I need to supply some account credentials and what you would do in order to get access to those is that you would go into the mongodb.com.

C

B

C

Here, mongodb.com and you would create an account or if your organization has an account, you could use that and then you would get those uh details from um for the organization access and so there's something that's called api keys. You can generate a new one, so in this case I generated one for the ocp sandbox, which is somewhat which I'm using and um yeah, and then that's basically the details. You know that's the public key private key and I will also need the organization id which you can find here. Okay,.

B

Thank you very much.

C

Yeah also, when you pick this one it it does offer you you know mouseovers, so you can tell you where to get those okay.

D

A

C

So now we have kind of connected uh openshift with the with my my atlas account, and I would like to also add an actual atlas cluster here, we're actually going to use the the.

C

The the the mongodb atlas standard, let's say, operator the 0.2 version to to connect to the actual cluster. I I do have a cluster running already, but I do want to also show you that we can create a new database instance using this. So that's what I'm going to do and I'm going to use this provider account that I have defined and I'm going to create an instance which is called something.

C

Fruits focus, maybe and then also.

C

Focus project, something like that.

C

And now that's being created, the operator is telling atlas that it needs to do things, so we can also have a look. We now see that we have.

A

Can you increase a little bit the phone uh behind this, so we.

B

C

B

A

Like this and on the font sides, so oh yeah, yeah, okay, cool thanks, okay.

C

Yeah, so now we can see that we are, things are happening. There's now, this fruits workers project that we defined and we have this fruits forecast cluster being created, so yeah yeah.

B

C

Is all yeah- and this is all controlled from where between the the operator, and so you can also see.

B

Yeah, if you click on view database instances, what does it say, yeah.

C

So it's saying that the results are going to work response, so it's informed about what's happening. You can also see the the new um the new cluster that I've been creating. Now uh is also this visible here. So these are um basically the way that you would add. Those is that you can, you can do search and you can do search for resources and you can do atlas cluster. You can add that to your navigation.

C

So that's what I did and then you can see a little bit more detail and, for example, you can, you can actually see the status you can see that it's creating and the clusterread is false and yeah. You get a lot of details and it's continuously checking the status of it. Okay,.

C

So what I also want to do is I want to add that, on as it becomes available to my developer perspective,.

C

But I also don't want to do that before it's been completed because.

C

Also checking this one.

C

You will see that the at this point when it's still not ready, there is still not the connection string available here, so we need really needed to wait until it's done creating uh before we actually and.

B

Is there another place where we can see the status or is that the only way in the yellow.

C

I would expect that it should show here, and you should also be able to see yeah.

B

Okay, yeah that down there you can see yeah cool.

B

Yeah, so you get the feedback from um from the openshift ui as well.

C

Yeah exactly yeah yeah, that's really nice! So you don't uh strictly need to go in into atlas 2 to check. um You are informed directly from within openshift.

C

Just realizing.

D

B

We didn't take our coffee, be.

D

No coffee for me today bring out some coffee beans.

A

Oh, come on straight to the plan.

C

Yeah now it's a good time for coffee.

A

Nice um yeah that that's cool uh johannes.

D

A

And uh I was, I was saying I like I are you customized at all the web console, so you.

B

A

The your custom resource, basically the operator, is bringing those- and I wanted just to say that this operator has been already installed in this openshift cluster, so the new user can have it. But if you have your own openshift cluster somewhere, you can install it from operator hub and and you you will have the same uh experience.

B

Yeah and so now we can see that the clutter is ready. Yeah does it say status true, so that's good news.

C

Yes, that's good news.

B

C

C

We can actually get that one into our topology to add the instance resource, and one more thing that we need to do is that we need to make sure we can connect to it because by default there is no outgoing um well there's no yeah, no um way that you can connect to it, so it's very secure by default, but it also means that we need to do one more thing.

C

um So what we want to do is that we want to edit our atlas project and we want to add the no not that one. Let me get the right thing.

C

I want to add a project type access list so for this demo um I don't uh know exactly what the I, the outgoing eyepiece of the openshift cluster, is, so I'm actually going to allow access from anywhere.

B

So what you mean is that you can restrict the ip addresses that can connect to your database, just to make sure that not anyone can just come and have the right, password or whatever and and get connected.

C

Yeah, exactly so having the finding the right password is, is not easy to do, but but there's a there's, an extra layer of security that that we do have there. Okay,.

A

That's cool, so if you know beforehand the cedar block of the openshift cluster, you could use that uh just to restrict the x or the certain you know. Sometimes we do some uh segmentation of the cluster where there's the.

C

A

Zone or the worker zone? No, so you can elect a couple of nodes to be the database nodes and only the ip address of those nodes can access the the database. So it's uh it's powerful to control from the develop the user. No also the access to the to the cluster.

C

Yeah yeah, and what you can also do is that you can set defaults on the organization level so that you know each specific team wouldn't necessarily need to do that. But you know you can actually inform the operator about that um it. Could I didn't see actually that with the the the operator that I'm using right now, but I know that the openshift um that plus 1.0 version does allow that and anyway now we are set. We have the ipads list and we have our service here.

C

We also want to deploy our application, so we can do something with this and we're going to deploy a container image, and this is probably going to be a common view for for a lot of a lot of you see carcass spinning up, but it's it's beautiful. So let's do that.

C

And we're going to have the quarkus icon, because it's beautiful as well and we're just going to go with the defaults.

C

So now we have our quarkus container spinning up here and we're going to use the surface binding. You see this little uh arrow here we're going to use that one to to have to use to inject the service bindings. So you know username password these kind of things and it will then create this binding for us.

C

Then the container is going to restart based on this new information and yeah. We can go and have a look. So this is uh the default landing page, but we actually want to to see that actual uh backhand and that's a little over here.

C

So now we can add fruits and then why not a banana, it's let's say, yellow and you can save that and the platform is, is now saved there um and let's have a look if some something actually came into the database here, so we can look at what's inside here we see that we have traction created. We see that we have now the banana and yellow here so yeah really simple to to do that. um There's a lot of metrics here as well that you can see a lot of things you can enable.

C

This is a little bit limited on the fit here, uh but anyway, and there's also profiler that you could. uh You can do um well if you upgrade uh also the performance advisor which is going to recommend index and schema. So there's a lot of tools for you as a developer. To do things right, something that might not be that well known is that you can also use full text search on top of your data.

C

So you know, and for this specific application, where we have you know user input, it would be really nice to have kind of a search bar box or something you know as the list of fruits grows ever larger and you will be able to like auto complete the name and stuff like that. So that's something you can do with uh without us. That's back end.

B

Very nice, and so just one one question about the the magic that happened. uh What are the, how is the connection made between the the the application and the back end? I'm guessing it uses some environment variables somewhere, but how? How do we get those? And where does it come from.

C

Yeah, so it's it's, this uh service bindings, so it actually is uh impact is using a secret okay, that it is well it actually uh yeah this and we're gonna go away. So it has the the the host uh of the server it has a password and all that. So that's the.

B

C

B

Yeah yeah, so so this the secret has been created, automatically yeah, okay and then how does the application know like where to store? What's the the database instance name and stuff like that,.

C

Yeah, that's part of the secret, so there's a convention.

B

C

Need to do on the application side so basically.

B

The application.

C

Needs to support service bindings and by supporting service planning, it means that it's looking for a service binding, um and what do we call it.

B

C

I think not this annotation, it's looking for it as a file. I think in the file system.

B

Okay, yeah, okay, like config map.

C

A

C

Like a conflict map, yeah yeah.

A

In this case, this square goes up. I think it's using the extension to use uh the service binding that are.

B

A

Instance in springboot so also node.js, so you have to use this plugin or extension to use cert the kubernetes service binding in the quercus case. I think it's gonna.

C

A

In the application properties, you know the host.

B

A

Username password, but uh in the other programming languages it's gonna, be it's gonna, be the same.

B

D

I have a question about the the roles, so um at the beginning of this of this demo, you showed how you basically set up through the operators, the connections providing all the credentials, the organization etc, and then uh you made available to the topology, the actual connection, to specific databases. Now.

D

Is it something that a developer would do or the developer has at this point in a list of databases that they have access to, and uh and so they don't have to worry about about connection, they may not even know uh all the credentials or is that not not the model.

C

Yeah, so exactly so um in this specif, it's a little bit special setup in the sandbox, because uh you do want developers to be able to set up the organization. That's also why I picked this way of the demoing uh but, for example, the the way that you would use the the atlas 1.0 version. um Then you actually, you would have someone like the organization that you're you know, use this admin.

C

The openshift admin to to set up your atlas credentials for like an organization-wide atlas account and then those credentials will not be stored in in individual developer workspaces.

C

So you will not actually have a chance to add your own um atlas credentials and you will not have a access to them either, but you would have access to to see what clusters are available to you and that kind of thing.

C

Does that answer your question.

D

Yes, yes, yes, um and I was also wondering if there is a way to to basically monitor uh usage, or um you know how many developers are connected, whether it is a developer level, whatever um service account that is, is connecting to that to the database. uh Just to have, you know, do a bit of house housekeeping in a sense, I'm always thinking of large organizations with a number of developers uh effectively they're using a shared database.

D

um That's that's on the cloud. So um would they go through a ui or is there a way to to understand who is using what database for multi-shift.

C

um You could look at that at those cluster definitions and you can also look at the service bindings, so the service bindings themselves would be specific to every application. So, let's see if we can do a search for a service for service bindings- and I would expect so- let's have a look and this you know. Obviously, if I'm searching a developer view I'll just find what's very visible to me, and but if you are searching as an admin, you would be able to search across all of them and then I would like to see okay.

B

Yeah, what are? Are you looking for a role.

C

Yeah, I'm looking for the uh oh, no sorry, yeah, you're, right role. Binding is the wrong one. I was looking for the service binding.

B

Yeah yeah so just type in service, uh yeah, okay, okay,.

C

Yeah here we go okay, so there are a couple of different ones. I guess the operators just.

B

Select select, all of them and you'll see, which one has like an instance.

C

Yeah, okay: here we go so this one and uh yeah, so basically you would be able to see. I guess you would do something like this and you write as well, so you will find those that start with, for example, but this is just the name right, so I could give it a different name. Also, but yeah I mean you can see.

C

This is the one that it's it's creating right now, so that's gonna give you a good idea and then you can always look at the details as well, so you would be able to uh see if what we are targeting the best service connection and and actually wasn't that easy to to see. I would say.

C

D

Would say it's probably.

C

Easier? If because this one is quite general general concept, it's probably going to be easier if you're using the the atlas 1.0, because it's going to be more specific and there's probably going to be more specific annotations to mongodb. On those.

D

I guess I was wondering if, um because I saw those extensions to the ui and I saw the atlas database users and I was wondering if that was related to basically assigning different users to different teams or or uh or developers, just to see how how that's.

C

Yeah, so I think there's something uh this seems to be some kind of bug. I think with this one right now uh on the.

A

C

Right now it doesn't show up, but otherwise. Yes, that is definitely the intention to mirror the database user resource and since any application would need it's uh only this user. That would be. You know. That would be something that makes sense right to to look at okay.

C

Indeed, so I would also say that I mean um having many um developers using the same database. That's definitely something that does happen. uh I would say with atlas it's, uh it happens, a bit less.

C

uh It depends on on how what kind of application you're building in some situations, you do absolutely want to to connect to the same database, and then this would be a great way, for example, looking at the database users to find out who's accessing, but otherwise you know considering there is a fit here right which allows you to start very small.

C

uh So even if you have just uh a small thing, you can you definitely have your own atlas cluster and you know, as you scale up and grow bigger you can, you can grow, it can go with you um but yeah. If you really, you really want to to get access to the same data, then yeah. This would be a great way to track that.

B

So, just uh so a few questions here, um so we saw that we are able to create instances from uh from here. uh Can we also like delete things from openshift and it gets reflected in in the cloud instance.

C

B

C

Correct- and I think this is uh an excellent exercise, so let's, let's do that? Let's do that, so we're gonna. You can see what I just did there. I was going into the atlas cluster resource here and then it's managed by this roots, carcass mongodb modulus resource, and then we have this database uh service, uh something I actually don't know what ddsi stands for, but it's something like that. It's also its own resource.

B

Yeah davy accessory systems.

C

Instances yeah so then I can go on this one. I can basically delete this.

B

C

And we can have a look also on the cluster side. It's immediately shutting down and cleaning up.

B

Okay, so it's like uh by uh directional uh synchronization and if we do create something in here directly, can we set it up on openshift like without using the automated uh integration as well so say, for instance, somebody has configured a cluster for you with some specific uh configs.

C

Yeah, so we saw before that when we started that this kind of uh cluster zero was showing up, and I had that created before so um so that's something we can. We can actually look at that uh already now. So if I go over here and we deleted the the.

C

Yeah one of the fruits corpus one, but the class. Well, it's uh still being deleted uh and this one is uh is uh still available right. So that was the one we saw.

B

C

Yeah, I wonder actually what's gonna happen now we still have this connection. We do need to clean up things in order. That's always a bit of a challenge with all these things. It's it's easier if you delete them in order, uh but we'll see.

B

D

A

D

It's fine. You.

B

Don't have you don't have to demo it it's just.

A

If, if we can like, have you seen that when he deleted the the binding, the app has been uh restarted,.

B

A

So removing the secret from the from the app.

B

A

uh Yeah, that's cool joanna said uh as jafar was saying if uh a cluster has been already created, you've shown that you can connect from there. uh What about the ink cluster experience? If I want my now, we we talked about you know atlas, which is a assass, but what about? If I want to mongodb in my openshift cluster disconnected, is that possible.

C

Yeah exactly it is, uh it is possible. So that's uh you have a couple of different ways of of doing that, so there is uh and the community operator. So that's also something that's available um to you um and that's a way to speeding up, let's say um yeah mongodb community instances, and then you have the the moodyb enterprise event operator, which is uh also about spinning up mongodb instances.

C

But then you have them automatically configured to to sync up with an ops manager instance, which can also run in kubernetes in openshift, and um then you get a lot of the tooling that I showed you really quick.

C

I showed you really quickly, um tooling, in atlas things like monitoring things like uh actually didn't, show you backup, but that is one of those things that are available uh profiler and the performance advisor tools like that which, which you know are, are connected directly automatically for you and and then uh then you have the assets operator so, but that does operate you that's you know uh connecting with uh with movie battles, uh databases service outside of photoship, but those two to the left would be what's uh available inside of uh openshift cool and another aspect with the with the ops manager part.

C

There is that if, if the customer already is running um openshift with ops manager in vms, for example- and they already have ops manager set up outside of uh outside of openshift, they can actually configure the the enterprise kubernetes operator to point to that ops manager as well. So what what that means for an admin administrator means that um they will be able to see not just uh the the mongodb instances that are created, let's say vms, outside of openshift, but they will also see um the movie instances created inside of photoshop popping up.

C

You know registering back with ops manager.

A

C

Yeah, I think this is you know. This is really a great mix of you know a kind of a devops way of working where you have developers able to spin up clusters and you know work quickly, but you still have all the operations tooling, and you still have the your let's say your traditional database team, for example, having insight into how developers are using the database.

C

So you know you still have a way for them to to give advice, or you know, if and and help out with best practice and and an easy way if, if the developers are, for example, calling the database team and saying hey, you know, there's something strange: that's not working, uh can you have a look and they can easily use the a great operations tool, uh but they, maybe you already know to to get a great level of insight directly from yeah the tools that they are familiar with.

C

And also for reporting, if you need to report back on the mobility usage, for example, you know having the individual mongodb databases basically registering back and that's really a neat feature.

C

A

uh So there's uh you.

C

Know plenty of options. I.

A

Think for a developer, I don't know what you think, jafar andrea, for a developer. To me it looks very, very easy. You know to get started with the sandbox and then you deploy your app uh container image or uh you can even start.

C

From the git repository.

A

And it will build, you know the the container. Then you create a cluster in atlas from here directly, so you don't have to go to auto security. It will it can done it can. It can be done automatically from to me. That looks very, very easy. What do you think about.

D

That's what I was thinking that in the case of a for developers, uh you you want to get them to not worry about resources on the local openshift or uh you're, not trying to trying to set up indexes or what I don't. I don't know what it entails to do: a mongodb to set it up in a way that it works for multiple developers, and uh that goes away immediately.

D

If you just work with the with atlas at that point, because uh they don't even apparently they don't even have to get out of the of the of the ui in uh in operation. To do to do that. uh Everything is also have contained. Of course, if somebody has set up the credentials before it's, even it's even easier uh that way. um I I think it's uh it's quite neat, actually that that we're able to do that so get developers to worry about the applications and uh and not too much about the infrastructure.

B

Yeah, so I think this is really nice uh it. It makes it very easy for for developers to get started, and so I I browsed abroad, the mongodb website, and I saw that there are some getting started, samples and uh documentation on how how to get started, and I was wondering if there are some like sample repos, that you can tell us about or demonstration applications that you guys use.

B

So you know somebody who wants to experiment with it has something that is like ready to use, or you know just for if you can share with us few uh few links uh that are useful for for someone who wants to get started. uh Yeah.

C

I think I mean I was actually I changed this a little bit there is. um I made up this uh basically for these instructions, for this.

C

For this, what was did would come up into this show lord openshift atlas public repo, and we have a little bit of instructions around in this case you know deploying the uh the carcass application and basically the steps that we.

B

C

uh Yeah that we would need to do in order to set that up so.

B

C

One example and.

B

And the the next thing is that we need to turn that into a quick start, uh natalie within yes, directly yeah.

A

Yeah yeah, oh you mean yeah, the quick start just say: johannes. We have an open, shifted. uh Those internal.

C

A

Maybe that could be a quick start on sandbox, because it's really it's really neat. No, it's! It can really.

C

Help on creating your first.

A

Road up based on would be atlas app that way.

B

Yeah, it's like very guided instructions on what you need to do and then it makes it even easier to consume or try. I would say the service and understand how it works.

A

Yeah, well, let's do it in a while yeah I'm gonna put the link of a web, a blog post uh that basically is doing a walkthrough. uh I think johannes did a better flow in this in the blog post. It's uh it's a little bit uh more more instruction. I think johann has shown, as shown has a very nice flaw on how to create a cluster and connect them, so um the blog post. It's uh it's a it's, basically a walkthrough, uh but uh if we do the quick start, it's gonna be much better.

A

D

I do have another question and again from the developer's perspective, so uh from what I had read about wanting to be.

D

There was also a search service that, if you have to do text search across the documents and I'll go through indexing, uh is it something that's already available through the operator or uh you have to connect to it in a different way?.

C

Right, uh I think that is uh that it wasn't in 0.2, so you needed to do something there and it might be on 1.0. I actually don't know okay, but otherwise it's it's something that you it's it's. Basically, it's a special type of index that you do want to control through the api, so it would make sense for the operator to control uh full text search indexing on on atlas um but yeah. Actually I don't know if that resources is already available.

C

I know that it's it. It is available uh on the uh let's say api, uh but yeah. I don't know if it's in the operator yet.

D

I guess you would still have to have your credentials and all that to to access the search and the service, but yeah. Okay. Thank you. Correct.

C

Correct so it's basically accessing. If you want to make a query, it will take search. It's basically you you would. You would connect through your database connection and make that search so yeah. You still know those credentials and in order to actually activate a search index, then you would need to use the api and you'd need to authenticate for the api as well, and but that's you know what we, what we set up in terms of that the movie atlas database provider.

B

Thank you so so I I do have a different uh type of question, so we we've had like technical insights, etc, and uh so I I see there's a building uh tab in there and I was wondering like from a, I would say, yeah building perspective when you are in openshift using this service uh and not using the free tier.

B

uh uh Is there anything that tells you about, like uh you know how how much uh the instances are costing or or how much they are going to cause and things like that when you are starting to create the the resources?

B

There's such type of information so, and also is there some type of control over over that? So you don't get like overly excessive bills and stuff like that.

C

Yeah, so you can, I know that there's been some yeah, you can set up alerts.

C

And that's on the organization level and that's.

A

C

Set up over there and then, if you want to uh you, know if you want to get an a feeling for the pricing there, there is a calculator here. Basically so when you provision a cluster of uh something other other than the fit here so now I've, I'm not picking the feature. Now, I'm picking the dedicated one and.

B

C

And yeah, in this case uh yeah this kind of instance size here yeah and it's yeah, giving you a calculation of that. So there is an api for that. I don't think that's integrated into openshift per se. So if let's say, if you're writing a config map, it's like in your text editor, it doesn't show up the price right.

C

Well, I guess that would I don't know how you do that, but um yeah it would show up here. I guess I guess maybe longer term it could be an interesting development that when you, if you're in the open shift, let's say cluster provisioning and you guess yeah.

A

C

Would be possible but uh yeah, I I don't think uh I I don't know how that's gonna play out, but I'm not gonna say anything about that.

C

Basically, for now, I would definitely if you wanna, uh if, if you're planning to spin up a certain size, I would recommend that you go actually into your project and you use this button here and you you.

B

And then connect it to openshift yeah, okay,.

C

Yeah you connected yeah or you provision it from within openshift. uh You can also do that. You don't actually need to create that one specifically, but you can figure out the kind of parameters right so um how they are going to impact right so in terms of instant size, ram, size, storage and all of that then you'll have an idea and then you can go back into to atlas and and set those um specifications up and permission from there yeah from upshift. If you'd like.

B

Cool. Thank you very much.

A

B

Cool very cool so.

A

uh So the people at the watching us will find in this repo the demo floor right, johannes. um So if they want to redo the demo again.

C

Yeah, so that's part of a pull request at this time, but it will be available there. Yeah, okay,.

A

As soon as joel wake ups uh candace and hello, joel thanks for also helping us in the in the moon, so uh he said that maybe he would join, but I think it's 4 00 a.m. There no.

C

Yeah, it is it's a bit early uh so but yeah I had great help for from joel and he I think at some point we'll we'll uh try to make a session with with joel as well, maybe for for the us time zones and uh hopefully as well. We can get back to you and have a session with the atlas 1.0 operator as as becomes available on the sandbox, because I think that the experience is going to be uh yeah really interesting as well to look into.

A

Yeah cool, um that's pretty cool, and uh just to recap, johannes, so developer. People are three way to deploy mongodb on openshift three formal way right, one is the mongodb community operator. So basically you can go to the this is useful when you wanna just try stuff in cluster. No, we just to consume the mongodb database in cluster and and then I put the link in the chat.

A

If you are interested on doing that, and and but if you need the enterprise support uh or if you need things like ops manager, you can install mongodb enterprise in uh in your openshift cluster right.

C

Yeah yeah- and I want to mention specifically why you like the reasons why you would want that, so one way is to get an overview of your mongodb instances from a database perspective.

C

The second one is to get monitoring and metrics to a detailed level to understand, what's happening inside of those instances as well, and the third thing is that it allows you to do backup and point-in-time restore. This is something that many, let's say, enterprise-grade applications do require. You know a way to not lose data um and yeah, that's not uh offered by the community operator and the the last thing is the database level encryption, that's also offered uh by ops manager and and the enterprise deployments cool.

A

Yeah, that's that's important if you think about a security stack where you know, for instance, in openshift, there's uh the operating system, the encryption of the operating system with looks then there's the sc linux context on the host on the containers and and then you can also encrypt the database in the encrypted file system. uh That's a very secure stack, I can't say uh but the third option. This is what we have seen today.

A

I know johannes the mongodb atlas operator, which is basically uh the way to consume databases uh as a service, and you can- uh and it's it's easy that we have seen it's easy. We have vienna an operator because there's a an abstraction like we have custom resource definition. We have apis to control this remote database from the openshift cluster and oh, it looks like joel is up so good morning. Bro, I'm happy to do a session u.s hours, hello, joel! uh Thank you for joining us, uh it's at 5am!

A

uh Yes, so we uh you understand, as I recap, this is what we, what what is the after offer today. So do you have any preference on this tree according to what we say we have seen today.

C

If you're asking me, I think you know definitely the atlas part, because it gives you uh you know, in addition to to just deploying a database in inside of the cluster. It gives you a lot of um optimizations that move down for the availability and the smooth operating of of mongodb. It's basically like it's kind of the promise of the operator that it's it's actually operating mongodb for you, but it's it's on a different level. You know with it when mongdb the you know, the actual mongodb company is operating your database for you right.

C

Then, then operator is kind of a good step up, but actually having the mongodb operating your database. For you, that's definitely the best place to be.

A

Yeah, so you are delegating this complexity uh to to assass service right right, whereas in the in cluster, you are delegating to the complexity, to the logic uh inside kubernetes. So if you want a full kubernetes experience, the ink operator, the enterprise, otherwise the atlas, it looks like the easiest part, uh also that you can scale up and about the scaling gap.

A

Johannes to do you, so I imagine you there's a different tires, not on on atlas, um but if, let's say we have an hybrid cloud apps that is spin up across multiple openshift cluster geographically distributed.

A

Do you have any blueprint architecture or recommendation reference on how to support such gel replicated application with atlas.

C

Yeah, so so atlas already offers a lot of that out of the box. So I'm glad that you asked so let's say that we go here and we want to create something and we want to set up, for example, a global cluster, and we want to enable global rights across the world. And then you could pick this kind of template here where you have.

C

The clusters set up in the the west u.s, the east u.s, the south america in emea, and you have uh in in impact north from tokyo and singapore, and so you can cover the the world and then, if you want, you can actually configure those to be multi-cloud as well. So you can actually say that maybe one of these actually, I would like them to be on gcp, for example, and I would like that to be in yeah the north virginia.

C

For example, uh you can also set those up to be cross cloud, and all of that so um and then the way that we we do this is that um we we can create one global cluster like this with a single endpoint, so your applications are going to know.

C

You know basically one address and then and behind that address it's going to resolve all of the the hosts and based on heuristics the driver is able to determine which, which cluster is the closest and that's the one it's going to talk to so from an application perspective.

C

It's really easy to set up a cluster like that, because you have one connection string that you would one basically one conflict map that you would push into your set of applications and then basically, the driver and atlas who figure out what is the best host to connect to.

A

Nice interesting.

B

Yeah, that's really really awesome.

B

That's probably one of the features that is much much easier to get set up as a service than if you were to try to figure that on your own, even if you are using operators and stuff like that, that's I would say some very, like that's some different kind of uh automation and and uh administration as a service.

C

Yeah and and actually you are able to to configure this uh kind of complex environment as well inside of the operator, so you could, you could actually have uh the operator manage also this kind of complex environment. um If you like, okay,.

A

Yeah, that's interesting. If you want to have an in-cluster strategy where the cluster are geo-replicated, you have multiple clusters. You want to have the in-cluster strategy because you want to have it on that in the data center or for any other reason. Oh.

B

And that's that yeah yeah- sorry, I was gonna ask: can this? Can this be mixed with like an on-prem instance as well, like uh you have multiple uh geo-replicated instances on those hosted services and one or two instances are like running on openshift or or even not on openshift with on-prem somewhere, and they can be as well part of a globally managed cluster? Is that something that can be done as well or.

C

So the atlas operator- well, the atlas in general- is- is really talking only to the cloud businesses right. So this uh the atlas, part of uh what yeah we have just saw that it's not aware about things that are living outside of of of the cloud.

C

But if you are interested in that kind of hybrid setup, the this kind of enterprise kidnaps operator with with ops manager is capable of that right. So that's that would be one way of doing that kind of setup where you still get a lot of the automation that that's available in atlas. But you get full flexibility on exactly how you want to deploy things.

B

A

And yeah folks, we are at the end of the session, but before closing up I would like to put some reminder. um So one of their reminder is that we have an openshift commons gathering happening in dublin if you are around and if you would like to hear again about rhoda and mongodb atlas. There's one speech from our colleague uh at. Let me check the timeline.

A

um So the speech is from banda simplifying database cloud services. We wrote is gonna, do a demo uh like like today using uh rhoda and openshi and mongodb atlas. So if you, if you like to attend it's uh it's cool event and other um for today, jafar, I think in the shadow we have the level up for today happening.

B

Yes, yes, so uh this afternoon, and uh so that's a, I would say, a good continuation because we will be covering uh caucus a little bit more in depth with uh danielle, oh uh which so I'm sure the session is going to be very interesting and very informative.

B

So uh yeah, don't hesitate to join us this afternoon on the level of power show.

A

Yeah cool openshift tv uh level power at 3 pm chest. We will come back next wednesday to talk with the red dot hack, fest office hour, and I take the opportunity to thanks johannes for being our special guest. Today. It was a pleasure to see you again, johannes. I hope to see you again in person soon and have a coffee or anything you like in sweden. What is the most uh preferred? Hot beverage.

C

Well, coffee actually is a super popular drink, so also coffee, yeah,.

A

Coffee definitely coffee. Let's welcome you together. Thank you, johannes, for joining thanks, joelle for waking up so early, and I hope we can make the same uh edition in uh north america jafar uh uh the same um uh demo the same in the same show in the in that um thanks uh uh thanks uh jafar, for joining thanks uh andrea for joining. uh I put the link in the chat of the our schedule today on openshifttv.

A

uh Thank you, everyone and talk soon.

D

A

D

Week with our iot folks of the red hat fest,.

A

Yes, yes, see you next week, wednesday 10 a.m, ciao bye.

B

Thank you. Everyone bye.

B