Just to give context:
I am planning to use Terraform to bring up new separate environments with ec2 machines, elb etc. and then maintaining configuration as well.
Doing that with terraform and using AWS provider sounds fairly simple.
Problem 1:
While launching those instances I want to install few packages etc. so that when Terraform launches the instances (servers) things/ apps should be up and running.
Assuming the above is up and running:
Problem 2:
How do I deploy new code on the servers in this environment launched by Terraform?
Should I use for eg. ansible playbooks/chef recipes/puppet manifests for that? or Terraform gives some other options/ways?
Brief answers:
Problem 1: While launching those instances I want to install few packages etc. so that when Terraform launches the instances (servers) things/ apps should be up and running.
A couple of options:
Create an AMI of your instance with the installed packages and specify that in the resource.
Use the user data script to install the packages that you need when the instance starts.
Use ansible playbooks/chef recipes/puppet to install packages once the instance is running (e.g. creating an opsworks stack with terraform)
Problem 2: How do I deploy new code on the servers in this environment
launched by Terraform? Should I use for eg. ansible playbooks/chef
recipes/puppet manifests for that? or Terraform gives some other
options/ways?
Not the intended use case for terraform, use other tools like jenkins or aws services like codepipeline or codedeploy. Ansible/chef/puppet can also help (e.g. with opsworks)
Related
We are using OpenShift. I have a confusion between buildconfig file vs jenkinsfile. Do we need both of them or one is sufficient. I have seen examples where in jenkinsfile docker build is defined using buildconfig file. In some cases buildconfig file is using jenkinsfile as the build strategy. Can some one please clarify on this
BuildConfig is the base type for all builds, there are different build strategies that can be used in a build config, by running oc explain buildconfig.spec.strategy you can see them all. If you want to do a docker build you use the dockerStrategy, if you want to build from source code using source2image you specify the sourceStrategy.
Sometimes you have more complex needs than simply running a build with an output image, let's say you want to run the build, wait for that image to be deployed to some environment and then run some automated GUI tests. In this case you need a pipeline. If you want to trigger and configure this pipeline from the OpenShift Web Console you would use the jenkinsPipelineStrategy in your BuildConfig. In the OpenShift 3.x web console such BuildConfigs are presented as Pipelines and not Builds even though they are all really BuildConfigs.
Any BuildConfig with the jenkinsPipelineStrategy will be executed by the Jenkins Build Server running inside the project. That Jenkins instance could also have other pipelines that are not mapped or visible in the OpenShift Web Console, there does not need to be a BuildConfig for every Jenkinsfile if you don't see the benefit of them appearing in the OpenShift Web Console.
The difference of running builds inside a Jenkinsfile and a BuildConfig with some non-jenkinsfile-strategy is that the build is actually executed inside the jenkins build agent rather than a normal OpenShift build pod.
At our company we utilize a combination of jenkinsFile pipelines and BuildConfigs with the sourceStrategy. Instead of running builds in our Jenkinsfile pipelines directly inside the Jenkins build agent we let the pipeline call the OpenShift API and tell it to execute the BuildConfig with sourceStrategy. So basically we still use s2i for building the images but the Jenkinsfile as our CI/CD pipeline engine. You can find some examples of this at https://github.com/openshift/jenkins-client-plugin.
I want to separate development and production environment variables for my shadow-cljs which is running in conjunction with a lein app through the Luminus template. My production environment is a docker container running on heroku.
It should work in a way that I can import a map, say config, and so that I can access the keys using (:some-key config), or something similar to this.
There is built-in support for separating release and dev configuration.
https://shadow-cljs.github.io/docs/UsersGuide.html#_release_specific_vs_development_configuration
The "import a map and access by key" you are asking for is not supported by shadow-cljs and would be something a library would provide instead.
I also do not recommend using environment variables to configure a build.
Also note that shadow-cljs configuration is about build time. If you want to access the "environment" at runtime you do not make it part of the build at all. Say you create a :node-script build running in node. You can just access js/process.env.SOME_ENV at runtime via normal code.
I have a web (online calculator for an example) which developed by my fellow tem members. Now they want to deploy in PCF using manifests.
Languages used : python, php and javascipt.
I gone through the docs about pcf with manifest.yml
In that I don't have any idea about services and env.
What is that services and how can I find the services for the above project and also how can I find the environment variables?
And tell whether these fields are mandatory to run the project in PCF.
To your original question:
What is that services and how can I find the services for the above project and also how can I find the environment variables? And tell whether these fields are mandatory to run the project in pcf.
Does your app require any services to run? Services would be things like a database or message queue. If it does not, then you do not need to specify any services in your manifest. They are optional.
Similarly, for environment variables, you would only need to set them if they are required to configure your application. Otherwise, just omit that section of your manifest.
At the end of the day, you should talk with whomever developed the application or read the documentation they produce as that's the only way to know what services or environment variables are required.
In regards to your additional questions:
1)And also I have one more query...like in our application we used python ok! In that we use lots of pacakages say pandas,numpy,scipy and so on...how can I import all the libraries into the PCF ??? Buildpacks will contain version only right?
Correct. The buildpack only includes Python itself. Your dependencies either need to be installed or vendored. To do this for Python, you need to include a requirements.txt file. The buildpack will see this and use pip to install your dependencies.
See the docs for the Python buildpack which explains this in more detail: https://docs.cloudfoundry.org/buildpacks/python/index.html#pushing_apps
2)And also tell me what will be the path for my app name if Java I can enclose jar files
For Java apps, you need to push compiled code. That means, you need to run something like mvn package or gradle assemble to build your executable JAR or WAR file. This should be a self contained file that has everything necessary to run your app, compile class files, config, and all dependent JARs.
You then run cf push -p path/to/my-app.jar (or WAR, whatever you build). The cf cli will take everything in the app and push it up to Cloud Foundry where the Java buildpack will install things like the JVM and possibly Tomcat so you app can run.
what should I do for application devloped using pyhton , JavaScript and php....
You can use multiple buildpacks. See the instructions here.
https://docs.cloudfoundry.org/buildpacks/use-multiple-buildpacks.html
In short, you can have as many buildpacks as you want. The last buildpack in the list is special because that is the buildpack which will set the start command for your application (although you can override this with cf push -c if necessary). The non-final buildpacks will run and simply install dependencies.
3) we were using postgresql how can I use this in pcf with my app
Run cf marketplace and see if there are any Postgres providers in your Marketplace. If there is one, you can just do a cf create-service <provider> <plan> <service name> and the foundation will create a database for you to use. You would then run a cf bind-service <app> <service name> to bind the service you create to your app. This will generate credentials and pass them along to your app when it starts. You app can then read the credentials out of VCAP_SERVICES and use them to make connections to the database.
See here for more details:
https://docs.cloudfoundry.org/devguide/services/application-binding.html
https://docs.cloudfoundry.org/devguide/deploy-apps/environment-variable.html#VCAP-SERVICES
My specific problem is that I would like to include the output of git rev-parse HEAD in a text file before deployment to ElasticBeanstalk. I believe the best way to solve it is to hook into the eb deploy command. I don't want to wrap eb deploy in a script.
I know .ebextensions defines commands that are executed on the ec2 instance during different phases of the deployment. Is there a similar utility for the client side?
I am migrating from DotCloud to Elastic Beanstalk.
Using DotCloud, they clearly explained how to set up Python Worker, and how to use supervisord.
Moving to Elastic Beanstalk, I am lost on how I could do that.
I have a script myworker.py and want to make sure it is always running. How?
Elastic Beanstalk is just a stack configuration tools over EC2, ELB and autoscaling.
One approach you can use, is create your own AMI, but since October last year, there is another approach that probably will be more suitable for your needs: ebextensions.
.ebextension is just a directory in your application, that get's detected once your application has been loaded by AWS.
Here is the full documentation: http://docs.aws.amazon.com/elasticbeanstalk/latest/dg/customize-containers.html
With Amazon Linux 2 you need to use the .platform folder to supply elastic beanstalk with installation scripts.
We recommend using platform hooks to run custom code on your environment instances. You can still use commands and container commands in .ebextensions configuration files, but they aren't as easy to work with. For example, writing command scripts inside a YAML file can be cumbersome and difficult to test.
So you should add a prebuild hook (example) into a .platform folder to install supervisor and a postdeploy hook (example) to restart supervisor after each deployment.
There is an ini file (example) used in the script; which is made for laravel specific.
Make sure that the .sh files from the .platform folder are executable before deploying your project:
$ chmod +x .platform/hooks/prebuild/*.sh
$ chmod +x .platform/hooks/postdeploy/*.sh