Cloud Native Computing Foundation Kubernetes Community Days (KCD) Chennai 2022, 30 Jul 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Workshop: Infra-as-Code & Terraform by Saravanan Gnanaguru

Description

Infrastructure as code (IaC in short) is one of the critical parts of DevOps and Terraform is one such tool to help spin-up infra as code. In this workshop we will discuss, IaC overview, Terraform introduction and demo. As part of Demo, we will create AWS infrastructure using terraform.

A

We are going to talk about uh what is an infrastructure code and how terraform is useful to uh implement infrastructure infrastructure as a code. So I think critique has pretty much covered. My introduction, like uh I, am working as a solution architect uh at infra cloud technologies, so yeah predominantly like you know.

A

I work on as a devops architect and like working on various cloud native initiatives in the company, and I am also an aws community builder and I would love to give back to the community via my technical blogs, and I also provide right answers to related to all devops and go related uh questions in stack overflow as well, and I also have my github in which I kind of provide a like workable poc solution related to infrastructure, code and devops areas, and I am on linkedin, uh twitter github as well as I have my profile in developtv as well, in which I write technical blogs, and I also ha have my website in which I publish my uh like work and everything is available there.

A

So this ppt should be available at the end of the session. I believe so. Let's move on to the uh the discussion. Part of it about infrastructure and perform so today this is the agenda uh kind of high-level agenda. Like that, I'm planning to cover.

A

We will go to go through about a iso overview on telephone introduction and then we'll discuss about various terraform concepts and terminologies, and then we will kind of jump on to the hands-on part of how to create aws resource creation using terraform and then we'll kind of have a small discussion about like perform best practices and we'll conclude, and then we'll have q a session actually. So this is the agenda for today, so uh when it yeah.

A

So let's start then so, basically, like uh first we'll talk about uh uh what is so infrastructure as a code is basically as a term uh uh like no mentions uh it is like no defining an infrastructure uh using a coding format. Okay, so so, basically like why we go for interest infrastructure as a code in uh traditional days like no, every infrastructure has been created manually for creating a simple uh like no server itself.

A

It was taking a week's time to like know, go through all the approvals and stuff and it was taking a longer time for a for deploying a single infrastructure right.

A

So that's when this kind of, in fact, infrastructure tools came into picture to defend the entire infrastructure, with all the network, activities and system configuration uh and like know as a code coding format, and it can be deployed uh into uh whether, like on a cloud provider or on an on-premises service, server, on-prem services to uh create infrastructure, basically, uh so so, basically like so so using this infrastructure code, it will helps to fix any configuration drift also in case if it is happening.

A

So if you are running a server in the sense, if someone accidentally deleted any uh kind of configuration, the sense that server may go down so using intra, we can make the configuration uh like no be in the desired state, always actually so that that configuration drift also can be resolved. And there are many infrastructure infrastructure as a core tools available now in devops space and right now, srv side. Actually.

A

So some of the tools I mentioned here, those are all like terraform, ansible chef, aws cloud formation, azure and gcp cloud deployment manager and the similar kind of tools. Basically- and there are also like no programming based infrastructure, also becoming famous like you know, terraforms udk and the awa cdk will be and all uh so. These are also some of the basic about new cluster code and when it comes to infrastructure code, there are two types we can split them. Actually one is mutable infrastructure as a code technologies and the other one is immutable.

A

So basically the major difference for mutable and immutable is something like you know in mutable infrastructure that infrastructure will be uh considered as a pits basically, but since, like you know, if, if in case I was talking about a configuration drift in one server or like you know few few number of server uh when it when that particular agent based infrastructure code runs on the particular uh instances or virtual machines that the configuration trip that that was missed in the previous like it was, it was deleted or something can be fixed.

A

Actually, okay, so that's the kind of mutable infrastructure that ansible and chef can be used to to implement such uh implementation and immutable is something like no, uh that infrastructure can be uh like you know, considered as a cattle, basically like. No, uh if, if, if there is one issue with uh some some configuration drift happens, this is you cannot fix that particular system alone, so we have to redeploy the entire infrastructure and so terraform and the aws conformation are such a kind of influence with the immutable infrastructure code technology. Examples actually.

A

So this is the. This is somehow a high level about infrastructure code as an overview and in this uh infrastructure code like no terraform, is kind of one of the very famous infrastructure technology in which.

A

Support for many cloud providers and on-premises providers basically, and it's kind of widely widely adapted uh influencer code technology. Now.

A

And we'll we'll see about a bit about terraform as an introduction, so terraform has been created by a company called hashicorp and they have uh multiple like other products as well, like vagrant, uh console or similar uh to other uh technologies and uh uh products. That's what they have uh and, as I mentioned earlier, terraform supports many uh cloud providers, aws azure, ibm cloud, google cloud, distillation and then on-prem side. They also support vmware kind of providers. Basically and telephone manages the external resources.

A

Whatever the cloud provider on on-prem provider, we are talking about right using the uh uh technology using the implementation called providers, basically like no uh terraform maintains terraform uh has a wider community support and it also have a like no library of functionalities for the for supporting their infrastructure resources. Basically, so they are, that is called the providers, uh and so ashitap maintains the extensive list of such official provider and can also integrate with the community developer providers. Actually so in case such for in some cases.

A

If there are some providers not there, and if that, if that particular instance or resources can be created via an api call right, then we can also like know, or anyone can also develop, that community based provider and it is called kind of a custom development of terraform actually like, but that is a kind of bit advanced topic and it's a kind of auto scope for this question session. Basically so yeah.

A

This is high level and, as I I mentioned earlier like this, is air one of the extensively used uh enforcer code tool and in the devops automation space. Actually, so that is the introduction of terraform.

A

So now, let's move on to the uh basic concepts of terraform, like you know how terraform works, and uh so so, basically like one uh take away from this session, could be uh for anyone who is starting to learn terraform or like no starting to wanted to build infrastructure using terraform. uh By using this session or by at the end of the session, they would be able to create the infrastructure, at least in the aws provider.

A

So that is the kind of uh like high level take away from this session, and it will be a very basic beginner session and will not cover any advanced concept. Basically, uh so so like no, let's dive into like terraform concepts. So like. Let's see like how terraform works so so terrafor in terraform, we we have like no, we would be defining the infrastructure using the hashicorp language and uh it will be having the respective provider calls to create infrastructure resources in the respective provider.

A

So how how it is internally working is like now for each cloud provider, for example, in aws, aws have sdks and api for accessing or managing their cloud resources internally, and that target api would be implemented. Using terraform provider that is kind of developed using gold actually, and whenever that particular functionality of uh running the creation of resources running on a specific provider, it will be interacting with the provider ap and it will be creating the resource uh actually, and uh so so, basically, like confirm, consists of three stages.

A

First, we will be like no designing infrastructure and then writing it as a code, and then we will be doing a plan of like know how that infrastructure, indicative instructions are going to be in the respective cloud provider and then once if you are con, it will show like you know anyway, like I will be showing some examples of all this, like you know how to write and how to do plan and apply and all, and once after that plan.

A

If you are kind of satisfied, then we can move on to uh like apply that particular infrastructure onto the cloud provider and basically like terraform, has an extensive support uh for uh techno uh community, and it is also having a detailed documentation for everything. So in this each slide. Actually, I have provided links for each of the documentation references.

A

So anyone wants to debug on the terraform concepts they can just do like they can just access this clip and they can read through all of this in detail, actually uh that that is a very basic concept of how terraform works. Basically like it is underlying layer. It is interacting with the respective cloud providers apa to uh manage their resources. Actually.

A

So that's the high level idea for this slide and terraform has a uh executable uh called the executable name itself is terraform and we call it as a cla term and terraform using the transform cli only we would be uh kind of uh uh kind of uh it. Probably it is basically providing a various commands to manage our infrastructure and some of the commands I have provided in this slide, basically like.

A

No, it is having a command like terraform in it in which, whenever we are doing the term init after creating a hcl code for a particular resource definition, we would be running the init command, so it will be interacting with the cloud.

A

It will be looking for the what all the providers defined in the telephone file and they could be uh like spinning downloading them into the user's local location. Other than a validate command is something like a syntax checking of uh like no terraform code. Whatever we have written and it there is also another command called the terraform fmt like go away, fmt command. That would be kind of doing a formatting of the terraform code.

A

Whatever we have written and then once it has been confirmed, we can do a terraform plan in which, as I already mentioned, yeah, it would be showing indicative uh infrastructure, whatever we are going to create onto the respective provider.

A

Basically, and it will also show how what are all the changes that is going to make, whether it is going to add or it is going to change or it is going to destroy the resources as per the definition and then the apply command comes like once after we are fine with applying that infrastructure onto the uh cloud provider.

A

uh We can also uh do the apply. We can do the apply command so that that infrastructure will be deployed, and then, once after, like no everything is done like no uh destroy command is something like if you wanted to destroy the entire infrastructure that that is defined in the particular terraform code, actually, so that we can use the terraform destroy command actually and using this terraform uh sub command and help option.

A

We can also explore uh like no command line uh hello, like options for each of these particular resources, and then we can uh like know we can. We can come to know like about more about each of these uh commands. Actually so I'll move on to the next slide yeah. So, as I mentioned, like hashicorp, language is the one in which we will. We will be defining our infrastructure and it will. It will be having various blocks. Actually, we will also discuss in detail like what are all the blocks.

A

Actually, one such example is given here uh basically like it will be resource. Resources is a block. It will be starting with a block type. In this case, uh resource block has a aws vpc. We are going to create so the resource name defined in the like hashicorp module is basically it's called the aw cpc, and this is the uh like no identifier for the particular uh vpc component within the uh terraform code.

A

Actually, so this app ppc name would not be the name that would be created in terraform, but it would be the internal uh representation of particular resource name that we can say actually and then inside that we we would be having various argument, references in which the terraform has implemented the aws implementation, uh like exactly the same implementation of what squared arguments and what are the optional arguments.

A

So, according to our need, we need to define that argument references and we can also add tags actually for the particular resources so that it can be easily identified as a group of tags in the aws entity. So this is about, like you, know, a high level uh defining a hiccup language for defining a terraform code and, as I mentioned like no, I have also provided a document reference. So whoever wants it wanted to go through about this deep dive on this code.

A

They can go through that, since this session is more on a hands-on part, I'll be kind of uh running through this high level concept. Actually so so these are all the basic uh like components of terraform or like terminologies. Basically, so, as I mentioned that like not in terraform, you would be defining the resources. The resources is nothing but, as the name uh represent right. So the resources are the uh component of the power cloud, powered or on-prem provider in which it would be an instance, a virtual machine instance or like no in gcp.

A

It would be a computing instance, something like that right. So that would be the thing and the networking components also called like whatever that would be created in the cloud provider or like on-prem product is called resources and the provider name is the.

A

As I mentioned, like you know, provider is the one that would be the cloud provider or on-premises provider uh that that is kind of implement, implementing the underlying resource management apis so that it can be uh written in the terraform code and they can be like no uh used for deploying that particular uh resources on the provider and module is something like no. We it's a kind of modularizing to avoid the code redundancy of such resources.

A

Basically, so, for example, if we wanted to define a ec2 infrastructure in aws, we can define a module for ec2 infrastructure and that model can be reused multiple times for creating uh like environment, size infrastructure like or by by passing the different variables to the same code so that uh the code redundancy won't be there. So that's what we have kind of like explaining this one in a module and the provisioners are something like no.

A

In each cloud provider they would be providing a facility to run some initial commands whenever the vm is getting created right. So, basically, if in a database it's called user data and in uh in gcp it's called metadata and azure. Also, it's called the user data, so in in provisional we can define such kind of user data so that whenever the vm gets created, we can provide some basic commands uh to basically execute that can be executed whenever the vm is gets ready, after that it is getting spinned up actually in the provider.

A

So this this this kind of provider section, can be useful to install some basic packages that needed to be available whenever the vm is getting whenever it is like accessibility. So we can use this provisioner for that and terraform has various variable uh like implementation, so input variable is something like no kind of global variable. We can declare uh in which we can, uh like you know, pass past the default value and the default value whatever is defined in that uh input variable that can be overridden by the file called tfs file.

A

So, as I mentioned earlier like no, uh if you are having a module and if we wanted to overwrite override the values that was defined in the variable section, we can pass the tfs file for the particular infrastructure to pass our like know, values according to our need actually, and the output section is the one in which, like you know, after the vm, is getting spinned up or like no, whatever infrastructure resources getting pinned up right.

A

For example, if you wanted to print the s3 bucket domain name and all that would be known after only applying the infrastructure onto the aws, so we can print that value so that we can immediately access that particular aws bucket whenever that infrastructure is like completely creating, and similarly we also have a local values in which that scope of the value will be available only inside that particular telform file, and we can use this file by local values to define some of the intermediate values that is needed for creating your infrastructure.

A

So these are all some of the basic concepts, resources, provider, models, provisions and terraform variables. Actually.

A

Yeah, so next is some of the important concepts in terraform we have. uh They are actually a terraform state file and a terraform dependency log file. uh So whenever we are like no creating a plan right so uh or applying that particular uh infrastructure right, it will be creating a terraform state file uh in which it would be updating that that uh infrastructure resource names and ids and secrets everything that would be stored in the terreform state. So this state file would be used to know what is infrastructure that gets created in the cla provider?

A

Basically, so so you. So. This is one of the key component of infrastructure uh in terraform, so that it needs to be like no not exposed in to the outer world, because it has user sensitive data in it actually and whenever we are doing it that perform destroy our terraform plan whenever there is a change rate, so that change is also getting tracked inside the terraform state file.

A

So uh so, whenever we are destroying uh some electron infrastructure, it will be looking for a terraform state file and, uh and it will be uh like no uh changing or destroying the infrastructure according to the state actually and the dependency lock is something like a question constraint file in which we would be detaining the provider.

A

We have a provider block definition right, so the provider talk, we have a version tagging for for the providers and the terraform versioning and all, and what perform lock file will contain is basically it will contain the basic, uh like version, locking of the provider files uh or whatever all the providers defined in that uh format, so that, uh if, if, if, if we wanted to, uh if this is basically and not to change the infrastructure, sorry not to uh like know uh vary the provider version whenever the same code gets deployed in the other places right.

A

So that's why, like now, we need it. It is also important to like now checking this file as part of our source control, so that provided locking will be also be tracked actually in the uh source control, so basically yeah.

A

So this is about a high level about uh uh terraform uh infrastructure components, uh and so this is uh highly about the term concepts and uh terminologies of uh terraform, and now, let's uh try to uh deep dive into the demo part of terraform, basically in which I am going to cover like know how to install terraform and how to configure aws credentials and I'll also help you guys to uh create some infrastructure for a beginner level right. So that's what I am planning to do now.

A

So, basically, for this uh workshop, actually I have created a uh folder and uh let me check in all the module files there. uh So basically we can just go through about installing uh terraform, like those kind of steps here.

A

So, as I mentioned, like no for uh basic components of terraform is something like no for any cloud powder. If you wanted to create infrastructure, we first need to uh install the particular uh like no uh command line uh tool for the particular provider, uh so just to like know, configure the infrastructure. uh uh Related provider related uh authentication details uh into that particular system. Basically, and uh for like you know, for aws, there are multiple ways now being supported by terraform, so I'll just show you them.

A

uh So whenever we are planning to create a authentication for aws right, there are multiple ways we can provide the authentication details to terraform. Now in earlier versions of terraform, we have to exclusively define the provider configuration and the path of the authentication detail, but now terraform by default, understand it from the mechanism that is implemented by aws. So first it will look for authentication details in the parameters and then it will look for authentication in the environment variables.

A

Basically, the aws authentication would be the secret key access key, and if there are any tokens we needed we needed to define for the particular session type. So those kind of detail you we can get from our aws account and we can set those details in this one and the third one is a shared current sales file and there are few other methods mentioned so for this demo purpose, I would be using the shared credentials mechanism in which I have installed the aws cli in my local, uh using the method I have mentioned here.

A

So so basically, you can see here right. So basically, this is the method to install awcla command and we can use aws configure command to define our access key and secretly.

A

Since I have already added my access key in secret key, it is showing that value here, so whoever trying to create for the first time it will not show any value here. They have to paste their access key first and then secret key and next, and then these two values they can provide their own for this value. They can provide their preferred region of in which their their resources needs to be deployed. And fourth, one is the json uh like no format or like aml format, whatever it is, we can define.

A

So these are all some of the labels configuration and what happens when we do this right, so these details will be written in the aws file that is located in the home part. Actually, so we can see here right, so basically it would be.

A

Whatever we are adding as a credentials, it will be added here and what would be we are adding as a region and json that would be added in the config details. Actually, so just to confirm uh I'll be doing my uh just once after we configure this, we can just do a uh check off like whether it is working fine or not by doing aws ec2.

A

Let's try instances.

A

So since I don't have any instances in my aws account, it is not showing anything. So this is just confirming that this aws configuration works, fine actually, so that is the first part of uh like no uh second part of this uh aws configuration basically configuring the adabs credentials. So in this method, I I have configured aws, actually using aws, configure command and added my credentials, and then next is like installation of uh terraform.

A

uh So basically, for that, so terraform has an official documentation, and now it is supporting a packaged kind of mechanism to install our terraform code, basically so that we can just copy paste the command and install terraform actually in our local. It is also having rhl federal amazon, linux and homebrew, and a few other uh like operating system transfer to install since I'm using linux. I have installed a terraform and as part of like a preparation of the demo, and it is also ready in my system.

A

uh So all these details have been. uh I have documented clearly in this one. So whoever wants to start with exploring that like terraform, they can use this readme file to understand more about it. So that's about uh terraform like a basic basic setup of aws configuration and installation of terraform.

A

So so, as I mentioned uh once after install perform, if we type perform command, it will be showing that uh what are all the commands basically right, we can use the terraform and I have installed the latest version 1.2.2, so that can be checked using the transformation command as well. So these are all the things that I was talking about uh like no knowing about uh how to uh uh knowing about the each of the subcommands of terraform.

A

So so this is about like insulation of terraform and configuring adiabatic credentials. So any questions you can post in the chat or like know in the slack channel so that I can see through later and I'll just reply to that and let's now jump on to creating the first terraform code.

A

Actually, uh since I would, uh as I mentioned like terraform, has a detailed documentation for everything uh I would be like more or less referring to the official documentation of terraform to or create my like, whatever the demo part, I'm going to try now so basically like, as I mentioned earlier, uh so basically for terraform, you would need a provider definition and some of the basic components right. So, let's uh kind of deep dive into defining those things. Okay. So what I'll do I'll?

A

I'm planning to create one directory now, just for the demo purpose and what I am going to do, I will be creating the uh terraform hcl file, so the telephone, hey file would be called as main.tf. Actually so I would be creating that and just open this in the vs code.

A

I will be creating a main file.

A

So I have created a main file so in so whoever wants to create a terraform uh code right, so they can just do a google search like we are going to do a uh aws terraform uh like implementation right. So I will just go through this uh search stream called aws telephone provider.

A

It will take us to the hashicorp's official aws documentation in which we are kind of getting to know about how to implement terraform for perform uh version 13 and later and telephone 12 and later so so, basically like no, I would be since I have the latest version of terraform. I would be using this so, as you can see, this definition has like no terraform required provider section and the provider section for aws right, so I would be using it.

A

Basically, we need a provider section to define our.

A

Deploy our cloud actually deploy our resources and then what I would be doing, they would be doing it taking an example of creating an ec2 instance. Actually, okay, so let's do a search on aws, terraform.

A

It will take us to the documentation part of uh terraform like terraforms aws implementation, so in which we have like no examples uh being defined for defining our various aws infrastructure.

A

So what I am going to do, I have already uh let know so so if you wanted to use any of the examples uh having your they are mentioned here, you can also like use this uh uh code. uh Otherwise I have kind of defined a basic code for defining the aws infrastructure in the workshop module. So I will be using that code here uh so just to implement that right. So.

A

So we are using especially two, so I would be changing that this x2 region and then um instance, key name and all uh I'm not going to define now so so so. This is the bare minimum uh example. uh I can say for uh creating the perform actually that from code, so.

A

Let us run this code now. So, as I mentioned, we have provider and a resource definition has been declared here and let's run this code.

A

So so I would be running this terraform init command, so it would be downloading all the necessary packages required for this code to run actually uh so it would be running downloading them to my local.

A

So yeah, so now it has been like no telephone unit command is completed. So, as you can see, this init command shows. Let me explain like now what is getting showed in any command so basically like terraform.

A

As I mentioned earlier, so while doing doing the init command it will, it would be initializing the terraform back end actually in our local. uh Basically, uh that is a lock.hcl file which contains the provider uh like no version, locking and every detail that would be available in this uh terraform init command. uh Sorry log file, so telephone we need to be initializing the back end, basically in our uh local and so a little bit advanced concept. If I can talk about, there is also way to uh keep the back end.

A

Basically, so we can either to keep the terraform uh log file locally, or we can also store it in the s3 kind of like no remote storage, basically in which, in which time we have to use a terraform backend as a remote remote. Actually, so so that's the kind of high level information I wanted to say here, but for our implementation it is just initializing the backend in our local and downloading the necessary files. It is needed for for running the telephone code.

A

Actually so it is creating a terraform dot, reform folder, uh and you can see that it is downloading the terraform package provider package from hashicorp, and it is also like no marking what are all the versions that is being locked in that is getting created while doing the init command, actually, okay, so that is about indeed command once after this init is done.

A

We can now see like what do the plan command to see what happens so, basically, while doing the plan command, it looks for the credentials and the authentication details of the respective provider, and it will show us like no. What is the tentative infrastructure that is going to get created.

B

A

So this is what actually it is getting showed. Actually uh you can see like no, uh our so our resource. We are going to create this database instance and the name internally.

A

We have defined our sample ec2 and what are all the values we defined for like now, creating the infrastructure it we are kind of like no, it is getting showed here that ama details and those kind of thing, and it will also show, as I mentioned earlier right so one this was going to get added actually, so it is showing that one is added and zero change and the zero destroy right. So so that is what uh that tentative plan. It is getting showed and then next we will try to apply this.

A

Now, what are the tentative infrastructure dish? It was showing right, so it will be created in the respective cloud account of of mine right so that it is now confirming that. Are you going to apply this like no deployment or not? So I would be typing s now. It will start spinning up the instance in the airbus cloud account whatever I have.

A

So you can see like now. This sample instance is getting created so now in the meantime, I'll just go to that error based console and I'll show you that that instance would be getting created here.

A

So since it's it's getting created, it is in the pending state.

A

So you can see the progress here in the console itself that what are what all the resources getting created. So it will be showing that the progress in the uh console uh actually so now that the resource is added, so you can see that a play is complete and there is one resource is added right. So you can now see, come and see that, like console that that uh resource is added here, okay, so this is the very basic of like no creating an aws infrastructure uh like note in uh lecture terraform.

A

Actually so let's go to the code like whatever we had actually so yeah so.

A

So this is the code we had so basically, in our code we have defined provider terraform, uh sorry to provide aws, and we are we are in. We are indicating the terraform like package that it needs to be downloaded from the hashicorp aws module and if, if at all, if there are like no custom provider created by our own purpose, then we have to source that particular provider name here, and we have to provide the path of that particular provider path in our own github repository or whatever.

A

It is right, and we are mentioning that that the minimum version of this provider should be version 3 and they should not be exceeding more than three actually so what it is doing is like it is downloading the last package that was available in the version 3.75 that was seen in the a cloud init uh command, actually sorry telephony command and then um in the like this provider part. We are mentioning the region right and there is also few other variables that know.

A

If if we are managing multiple uh aws profiles right, we can also have a variable uh definition, argument definition for profile and we can also provide uh like the profile name so by the by default. Aws uh have the profile name as default. So if you wanted to uh like indicate that if it is something else right so by default, it will be creating the default parameter default profile only.

A

But if you have multiple profiles for creating infrastructure, like you know, production some other account or for various environments, various profile industries, then we have to mention the profile name accordingly. Here in this profile detail actually so now, then we can. We have this resource definition path, part where we have this uh aws instance resource defined, and then we have provided the ami.

A

So this is the ubuntu ami, and we are also mentioning that what is instance type for creating this a instance actually and then we are kind of adding tags, uh and this is the user data section. I was talking about right um in the particular part, so this script will be executed uh whenever that vm gets created uh so basically like. No, I will also have a detailed example of like know, creating apache server. That is a kind of larger example.

A

I have created for this demo, but whatever code I have showing here is anyone wanted to start with aws? They can simply have these three definitions just for their learning purpose and explore more according to that they can add more resources for part of it. So for that I can summarize that basically, they would need a a like.

A

No aws, credential and aws account taxes, and then they have to have a telephone cla command installed on their local, and then they can use terraform documentation uh to have the basic need of code, and then they can create information and hashicorp also have a portal called learn.hashicob.com. So these are all the details more or less covered. There are there also, but I wanted to provide for us for beginners benefit. I have.

A

I am sharing all these details actually, so this is about like no high level, and then these are all the code blocks that is very basic for anyone just getting started with terraform. Actually so this is about it for the live demo part actually under them. I will also uh show you the other parts of code, whatever I have created and uh let's also see like what are all the other topics. I I'm planning to cover in this demo.

A

Part basically like we have covered, like writing the terraform code in aws, and we also like made a walkthrough of creating an easy to example in our like live demo code right so now, and then I have also showed you like.

A

No, what are all what are all the commands that would be useful for like no uh getting started with uh terraform like the terraform unit command will be initializing the back end and would do the version locking in the log file under the terraform plan would be showing the tentative infrastructure and what are all the changes, destruction or like you know, additions that particular the code is made to use in the cloud provider and then therefore apply command will be actually applying that particular infrastructure on the cloud.

A

So now I have added one change in this file. Basically, I am. I have added a uh like no profile argument here, so let's try to do a plan now again and see like know what it is showing.

A

Let me do it here.

A

So you can see that uh it is reading the state from the state file telephone state file, and- uh and it is mentioning that so since we haven't changed anything related to infrastructure, it is showing that no changes are made, and one part I kind of listed to mention earlier was after doing apply. It is creating the state file right. So, as I explained in the introduction right in production part, it would be having the details of uh details regarding all the uh like resources associated with this instance id.

A

Basically, it is also having the ebs volume id. It is also having what is the vpc and subnet. This particular instance is getting deployed onto and it is using the state file to uh kind of.

A

Using the state file for understanding what all the changes made in this infrastructure actually.

A

So then we will try to destroy that infrastructure, so.

A

This is for, since this is a kind of demo, uh we also see like what what steps are involving in destroying an infrastructure so so for destroy. Also, it will be asking for confirmation.

A

So again, it will be using the terraform state file and see what all the resources that are associated with this particular run after form state, and it will show that changes, so it will be showing that plan one plan it is showing that one would be destroyed. Actually so then we can confirm that. Yes, then that particular instance, whatever we created that will be destroyed.

A

So you can now see in the console that that instance might be uh terminated, so it's now shutting down and later it will be moved on to the terminated state actually.

A

So so anyone can post any questions related to that and I'll be answering at the end of the session.

A

So, since that uh this infrastructure is getting destroyed, the state file doesn't have anything now, so that state would be moved into the backup stage. Actually so, and as I mentioned, uh this state file also can be stored in the s3 right. So basically, the advantage of storing the telephone state file in the sd packet is multiple developer environment right.

A

So basically, there is a situation that multiple provider would be running, like environment creation, right on the same code right so at that time, this lock and state file would be used to avoid that, and it will indicate that that the infrastructure creation is already in progress. Actually, so that's the reason like no in a multiple developer environment, uh it is better to maintain the state in the uh sd packet or similar kind of places.

A

So now that instance is getting uh instance got terminated right. So now, let's move on to few other examples. What I have diff uh like sorry, what I have explained here uh in this uh deck so basically like no, this is the.

A

This is the examples I wanted to explore more so basically like. I have an example for module, moduling modularization as well as s3, and one full example of creating apache web server. So let me show you that.

A

So, as I mentioned earlier, so this is the terraform uh like no demo code. I have created right, and this is the same code. I I try to. I try to run using the live demo ah as well, so.

A

So so like, as I mentioned earlier, like no, we are having a concept called modules right, so modules would be useful for modularizing the code or resource right and then use the.

B

A

Right so we need to keep this provider definitions in provider.tf and resource definition and those kind of definition in main.tf and similarly, output, tiers.

B

So that kind of modernization we need to do whenever we started. So since this is the demo, I haven't kind of created, separate files just to show what needed for creating a basic performance. I have to put everything in the same file.

A

But in in my other example of ec2 right, I have kind of modularized everything like now. I have. I have created separate files for each of the things actually so, basically like no main.tf only has the resource definition and output. I am kind of printing out the instance id as well as the public, ip, whatever that is getting created right and then provider.tf I would be having whatever what are all the provider definition I have had and then in variable.

A

Actually, so I'll just quickly try to apply this in the meantime, actually, so that we will see that.

A

Let's do a terraform in it since I have already run this uh telephone unit earlier in my local. It would be using that terraform log file uh in my local and it is showing that terraform is initialized and using the reusing, the previous version of log file right. So that's what it will show, but for the first time it will download everything that is needed for running this. And then I will do a transform plan.

A

So now it is showing that what are all things are you generatively created and I'll? Do it apply? So one uh quick uh like no thing on the option, part actually like no for terraform, apply and destroy. There is an option called auto approve. Actually, there is a caution that it should not be used in production kind of environment right just to avoid the prompt of uh having a s right whenever we are running apply command.

A

So for since this is a demo, I'm just adding the command for auto approving this right so that it will not ask for the prompt and it will start creating the infrastructure.

A

So it started creating infrastructure, actually, so let it let it uh let it create the infrastructure, and then we will come back to that later and uh let's move on to that modularization, uh like example. Basically, so what are what?

A

Whatever the uh ec2 instance creation code I have created for this demo right, I have made that same code as a module here right and in this module it will be having the replication of the same files, but uh while defining that like no using that uh particular uh creating the particular uh resource, what you have to do, we have to define a similar structure like no main uh like uh provider and those kind of things, since this is a demo, those files are not like added in this demo code right and what I have do.

A

What I have done is like now we are calling that particular module using the uh like a block called module block. Actually, in that model block, we would be adding a uh like no indicator, name that this is. This would be for the ec2 instances and in that particular module block. We are.

A

We are kind of sourcing, whatever the model we have created for creating the easy to instance right, so we are creating uh like now providing the model as a part like a path for that particular uh uh to get the code from there actually and then this model uh needs a region as a variable right, so we can pass that variable as well in this module. So let's try to uh like you know, apply this as well right, so so.

A

The earlier example of the exit instance was the completed uh whatever the code I showed here right so and uh yeah and and it was printing the instance ip and the public ip so yeah so and now we move on to that model part actually. So then yeah, let me go to that code and I'll do a terraform click here.

A

And in in this one you can see that it could be doing the initializing modules as well as initializing the back end right, because module is the one containing the actual implementation of the code right. So it would be doing the initialization of that as well. What I will do for the demo purpose, I will just remove that dot. Terraform implementation, whatever that was created already right and then I'll rerun that terraform unit command again.

A

So you can see right, it is initializing the module ec2 instances and that is available in the module directory and create instance directory, and then it is initializing. The back end, as I mentioned earlier, that logs will be created.

A

From the aws hashicorp module now it is kind of initialized. Then we will run terraform plan.

A

So now it is showing that one resource will be added here right, so we will do a apply here again.

A

So once after this is it which is getting applied, we'll just go through the code once again, just for the purpose of revisiting so now the model it is using the model code to create the easy to instance right. So that's what it is showing. So, as I mentioned, so we are sourcing the module path here and using the module block in the main.tf, and it is going to that create instance, modules territory and use this uh module uh for creating instance.

A

Actually, so, whatever the resource we are defining, so we can have multiple modules and multiple module definitions inside that, so we can have modules for vpc, subnet and everything, and we can just call it multiple times so that the code redundancy would be avoided. Actually, so that's the high level idea for this module, so it is now it is now created. So we can now see in the console that uh a particular instance would be created.

A

So since we have also ran that other example of uh ec2 right, so it was also showing that- and uh so now that created instance is ending with d1 f. So this is the instance that was that was created now, so this is about uh like no high level on how to convert our plane terraform code as modules actually so that is about the terraform model, and then I was talking about various sections like variables and stuff right.

A

So for that I have a sample code for uh like s3 packet creation right, so so this this code also like, uh like most simple, plain sd packet code that was available in terraform uh definition uh like you know that transform documentation. So basically we we would need. We would need to define uh like in main.tf.

A

We would need to define a three packet uh like no resource and then we would be providing the bucket name and then sp3 packet needing acl details right and then we we have to define that as well, and then we have to define the it's not mandatory, but for for the purpose of uh like know uh for the purpose of uh for the purpose of like demo purpose, I I have added this versioning block as well, so in this one also like I have kind of created a output file in which it will show the s3 packet domain name and sl packet id and then in provider section.

A

uh I have defined like no, what is the provider and what? What is that minimum telephone version required for this, and this required? This required version helps to uh like you know, keep uh indicate the user that this is the minimum, like no uh version of terraform. Anyone needs to use actually so that that can be used. And then, in variable section, we have this region file.

C

Yeah seven- and there are so many questions, so maybe we can start taking one by one.

A

Yeah, actually I haven't completed it. Okay, but.

C

Okay, then, then, let's complete and take it.

A

Yeah, just I just wanted to see that like no, uh whether it's active or not, so that's the reason I came back here. No, you.

C

Are in backstage, we have a lot of questions, no worries.

B

C

Okay: okay, okay, okay, only host you can see you are in the backstage, but.

B

C

Have a lot of attendees around 500 to 600 attendees.

A

Just wanted to check actually because I it looks like I was only talking, so I wanted to have a quick check on that: okay, so yeah yeah. So basically this s3 example. I was running through right.

A

So uh so one catch here right we talked about variables and we talked about output variables like in our in my high level definition I was showing about this variables block right and this. In this example, I have implemented something like a local variable mechanism right uh because in aws terraform, sorry in adwords, s3 packet name should be unique. It should not be created anywhere else right.

A

So what I am doing in this code is basically I am creating a random string and then- and I am attaching that string with the s3 packet name right, so that uh that name should be unique to this implementation alone right. So so that's the reason like I have defined a resource called random stream. These are these. These are some of the facility provided by terraform. I think that's that's available from uh later from the version 13 only it will not be available in earlier questions.

A

Actually, so you can check in the telephone documentation that in from which version this terraform random string is available, so for making that string like randomized and unique to our implementation. What I have done is, I have created a local section of block and in which I am defining the bucket name and in that packet name argument. I am passing that, like some common string sample s3 bucket, because this is a demo kind of thing we are creating- and I am attaching the uh like.

A

No, I am adding that string of particular random string that was getting created using this random string resource actually so that the sd packet name would be so unique for us. Actually. So that's the idea for creating this uh local. So this also shows the example of how we can make utilize of make you placing the facility of local variable in term and that bucket name would be referred in the s3 packet resource as a local dot bucket name right. So that's the high level and let's try to apply this resource as well.

A

So we have covered so far like creation of ec2 instance from the scratch right using telephone documentation and we have come across like what are all the basic components you get for that. And then we have modularized the same easy to instance, uh code uh under use the module block to create the instance and now we are seeing example of like local in the sd packet uh code. Actually so now now I'll move to that sd bucket.

A

So yeah so now I will be trying kind of doing the the usual procedure of running the tab for minute again like no. Since I have run it run that terraform block fill earlier. It is using the same log file and like now showing the message that reusing the log file and then run the telephone plan.

A

So it is showing that tentative infrastructure created what what is what it is going to create now and then I will do the apply and I'll use the auto updo command, as I mentioned earlier, be cautious, then using auto, approve in apply and destroy on the real time environments so yeah, since this is a demo code, demo purpose, I'm just adding it so what it will do.

A

First, it will do the random string creation right right and then it will be showing that so so now you can see right the terraform package, so the sd packet name should not have the other special characters. It should have only dot and hyphens right. So since uh it is creating a random alphanumeric characters right, so it is showing that only hyphens and characters are like no dots are allowed. So what we need to do, we need to change our code now right to uh accept only that right. So basically what I do I'll.

A

So these things can be explored in the documentation. I will show you that as well, so we are making. We are indicating that random string that there can be special characters in our uh random string and we can we need to. We can also specify what are all special characters that can be added in the uh string, basically right. So what I'm doing as per the sl packet naming convention, only dots and hyphens are allowed, so only I am giving that special characters in the sd packet randomization name right now.

A

I am I'm just saving this file and running the code.

A

Let's remove that state file because uh the state file contains the string right so I'll, just remove that and I'll redone. It.

A

So now we can see that the randomization stream created some random character right so that uh it doesn't contain, contain any other special characters, only the alphanumeric ones. So so that's the update we have done under the s3 packet started, creating.

A

Now we can see that the sd packet is created and the bucket name is sample right. So let's do a quick check on that sd packet using terraform, sorry aws s3cla command. Actually, so I will be just grabbing it for the name.

A

So it is having that s3 packet created in our aws account right, so I'll, just quickly run through this example once again, basically like in this example, we have created a three packet resource in aws account and we have defined various other resources as well like that are needed for creating sd packet.

A

So in this one we have acls and versioning also defined as a resources and for s3 packet name. We made we made the use of a local variable in uh like terraform, and then we have created the random string mechanism to create the string, and then we have created the bucket using that right.

A

So this is the high level about uh like using the locals in the terraform file, and it is also showing the example of uh like how it is changing things, and uh one other thing I wanted to show now is like quickly is a. Is that what what happens when we do a change right, so I'll, just change the string to disable the like. Now the versioning string has disabled right and then we will try to see like what happens now.

A

So I'll run the terraform plan. Once again,.

A

And it will show that one of the resource gets changed now that is actually the versioning resource.

A

Okay, so the name should be something else. So what we'll do we'll just quickly refer the documentation for aws three right aws, the perform? That's the launch name right.

A

So this is how like whoever wants to create work on terraform. If there are any facing any error, documentation should be the first place they are. They have to refer right.

A

So how to refer the documentation is basically, I will also cover that as well.

A

So we have this example usage right in which one in which we it will be showing that example of particular block actually and then there are various things like argument, reference and attribute reference, so argument, reference are the one that needs to be added inside that particular resource and then attribute reference is some of the outputs that is getting out of the particular resource, actually so for versioning. uh Basically, we'll just look for status.

A

Okay, so for that we have to uh recreate the entire resources. So so, let's move on to, like you, know, uh changing the uh like s3 packet acl, actually so that at least I can show the uh which one, like you know that changed basically like what happens when we are changing the environment right so just quickly, I'll.

A

One minute one minute just give a moment.

A

Sorry for this basically yeah.

A

Sorry, sorry for this, so basically what I'll do I'll just remove this acl itself? Actually because that's an optional parameter for the for the purpose of demo, so that we'll see like what happens now, because, as per the documentation, that aws uh is the packet acl option. You know that argument is an optional one. So we'll just try to like change that.

A

We will try to apply now.

A

So it is showing that, like anyway, that's by default a private bucket right, so so the default for s3 is a private one. So I think it will not like show any further changes. I think so, let's see that so one change it is showing now right as part of plan. It is showing that one is to change. So let's do that.

A

This is the plan it will show whenever there is a change in the particular infrastructure right. So in our case it is going to change this acl configuration, so it is showing that one is getting changed. Actually.

A

Okay, so I think uh some of the uh things are required in uh like mystery package team. So I think we will keep that as our electron separately and I'll show you some other example for, like the changing environment, but basically like this is how it will show whenever, like there is a change in the environment. Actually, so that's what and then it will apply.

A

And then the next example I wanted to show was.

A

uh Larger example of that creating apache server okay, so this is a kind of entertained example of creating a vpc subnet, and then it is also having an example of like no creating a apache server while doing the uh creating server instance itself, basically, so that uh it will, uh it will be like the mode actually so that it will try to apply that quickly. Okay, so let me go to that example of like showing that particular apache server example.

A

And for this one like no for this kind of implementation, I have added a detailed documentation as well in the workshop module, so that you guys can have a look into that and kind of make use of the documentation and know more about the code. Actually.

A

Okay, and so what I'll do is just destroy the like the bucket I created.

A

Go to that example of.

A

So this is the code actually aws. Tf code is the example in which you can see like various uh modules I have defined actually so, basically like no in main.tf, I have I'll show you here actually.

A

In this one you can see like no, we have defined the complete uh network, like that is required for spinning up a vm right, so we have defined aws, vpc and then the internet gateway for providing the internet access and then public subnet uh that is located inside the vpc and then a route table for that particular vpc and internet gateway. And then we are associating the route table with the subnet actually, and then we are coming to the part where we are defining the uh infrastructure uh for like for creating the uh instance.

A

Basically in this one I have my ama defined instance type defined and the instance key name also defined, and I am also associating the subnet id that that I have created in the above definition, and then I am also associating the security groups as well. So security group is uh nowadays by default. Aws doesn't allow uh like no public, uh like user access to ssh or uh outer world internet. Actually, so for that we have to have a special security groups needs to be created for allowing the ssh and http access.

A

So basically, uh what I have done is: I have created a security group file separately for defining the ssh taxes and the http access for the particular instance. Why are the subnet for other vpc security groups? So you can see like now that inbound is ingress and I have defined the ssh 22 connection and in grand indones, like http 80 also load and uh for outbound I am allowing all traffic actually. So this is the security group. I have defined. Let's quickly apply this code and I'll and we can see what happens so terraform.

B

A

So, basically, in this example, I made use of variables file in which I am kind of customizing various inputs. I I wanted to uh you know pass in into the particular infrastructure creation, so predominantly majority of things are kind of like can be used, but only this key name variable right. I have added using the tf vars file, so this is the tf pass implementation.

A

Initially I was talking about in the in the introduction part right. We can customize our own like variables and we can use this variable file alone to pass the customized variable inside our terraform code. Actually, so for that I am what I am doing, while applying the code, I need to pass the option called var file, in which I need to give the path of this variable file, actually aws dot, pf vars, in which it contains all the necessary variables, customized variables that is needed for this infrastructure creation.

A

Now I am applying this, so it will be using all the like, no variables that is needed for creating this infrastructure using the tf apps.

A

Since the lack of time, uh I couldn't uh like know show the change of like my s3 bucket options actually because that needs some exploration of options, uh sleep at a command creation. So basically, like that's the reason. Whenever we need to create a infrastructure, we need to keep ready with all our uh changes or definition that is needed to be created for defining or changing infrastructure, so that we can avoid this kind of error. And even if there is an error, then we can explore the reform documentation for fixing those errors.

A

Actually, so now we can see in this web web server example, it is started creating all the code in the meantime, I'll show you that what is available in the main.tf right as I was explaining like it, has vpc internet gateway, subnet and all the definitions. In the instance definition, you can see that I have defined the user data section right uh in user data, even though it accepts uh cell scripting and for windows kind of systems. It accepts powershell right.

A

uh That is the kind of aws implementation that is, that is made use in the terraform implementation here, in which I am kind of installing the apache simple apache server, uh using a four line, four lines of shell script. Actually, so what happens right whenever this instance is getting created?

A

This script will be executed and it will be kind of showed it will be having the apache installation done whenever the system is ready for accessing right. So now we can see that this instance is created. Now. What I am going to do now is, I am going to show that this script should be should have been executed inside the instance that got created right. So since I have configured a vpc subnet and using my instance accessing key right, that's a pem file. We need to provide what I'll do I'll.

A

Just just do a ssh and I'll do the.

A

This is a ubuntu server, so I'm using ubuntu username.

A

This might be felt too heavy for some beginners, but uh whenever you are planning to create a uh like, you know, uh instance or like a resource creation, you can just refer whatever I have like explained in this session, so that it would be easy for you to learn more on this right. So basically, this session as a told earlier, can be used as a basic thing to start learning as a beginner for creating terraform code.

A

Actually, so since uh it's asking for that, adding that known host, I will be adding it so now it will be going into my vm that got created right. So this is a vm that got created and it is, we are inside the vm. So, as I told earlier, it was creating a it was running a set of scripts right whenever the vm is created, so that can be checked in the cloud init log that would be available in the slash dialog part. Actually.

A

So this is just to show that this user data has been executed successfully.

A

So this is the user data file and it is having a set of instructions right, so you can see here right. So we are. We have given that, given the command of inside apache 2 that we can see it here, so it means that it runs this cloud like no user data section and run the script right for accessing this, uh for installing the apache inside the instance source.

A

So this is the kind of example that that shows that we can also use, may make use of this uh user user data section to um spin up like no whatever the beginning command that needs to be installed in the vms. Actually and now. I will also show the apache server that got created using this server.

A

So, since this is a public ip, we would be able to access it.

A

So this is the sample right, so this is the apaches uh default page that comes from the slashware log, slash, www, dot path actually, and so this is a kind of like no full level. Example. That shows how to create a entertained instance like with all the networking and stuff actually. So this is the high level I wanted to cover, and there are few other slides I wanted to quickly run through. So we are, we are pretty much done with the demo part actually and on a high level.

A

These are the best practices right uh use, the terraform bar files example. I showed you for customizing uh for sending the customize variable and the version control this that form state file for like no uh reverting back to the earlier versions or, like you know, for role for for rollback under managing the s3, that is, for the multiple developer environment, right, tf, state files, so so that it can be avoided for multiple turns of the same infra and then retrieve the state metadata from remote backend right.

A

So that is also similar to that and use shared module and isolate environment using uh variables and modules. Actually so these are some of the best practices and in this deck also, I have provided the kind of github reference for this particular best practice. Github page yeah- and we are pretty much covered. Everything and that's. It hope this session might be, might have been useful for people who are getting started with terraform right.