I have a python project and i want to deploy it on an AWS EC2 instance. My project has dependencies to other python libraries and uses programs installed on my machine. What are the alternatives to deploy my project on an AWS EC2 instance?
Further details : My project consist on a celery periodic task that uses ffmpeg and blender to create short videos.
I have checked elastic bean stalk but it seems it is tailored for web apps. I don't know if containerizing my project via docker is a good idea...
The manual way and the cheapest way to do it would be :
1- Launch a spot instance
2- git clone the project
3- Install the librairies via pip
4- Install all dependant programs
5- Launch periodic task
I am looking for a more automatic way to do it.
Thanks.
Beanstalk is certainly an option. You don't necessarily have to use it for web apps and you can configure all of the dependencies needed via .ebextensions.
Containerization is usually my go to strategy now. If you get it working within Docker locally then you have several deployment options and the whole thing gets much easier since you don't have to worry about setting up all the dependencies within the AWS instance.
Once you have it running in Docker you could use Beanstalk, ECS or CodeDeploy.
Related
I'm using venv for my python repo on github, and wanted to run the same code on 10+ ec2 instances (each instance will have a cronjob that just runs the same code on the same schedule)
Any recommendations on how to best achieve this + continue to make sure all instances get the latest release branches on github? I'd like to try and automate any configuration I need to do, so that I'm not doing this:
Create one ec2 instance, set up all the configurations I need, like download latest python version, etc. Then git clone, set up all the python packages I need using venv. Verify code works on this instance.
Repeat for remaining 10+ ec2 instances
Whenever someone releases a new master branch, I have to ssh into every ec2 instances, git pull to the correct branch, re-update any new configurations I need, repeat for all remaining 10+ ec2 instances.
Ideally I can just run some script that pushes everything that's needed to make the code work on all ec2 instances. I have little experience with this type of thing, but from reading around this is an approach I'm considering. Am I on the right track?:
Create a script I run to ssh into all my ec2 instances and git clone/update to correct branch
Use Docker to make sure all ec2 instances are set up properly so the python code works (Is this the right use-case for Docker?). Above script will run the necessary Docker commands
Similar thing with using venv and reading the requirements.txt file so all ec2 instances has the right python packages and versions
Depending on your app and requirements (is EC2 100% necessary?) I can recommend following:
Capistrano-like SSH deployments (https://github.com/dlapiduz/fabistrano) if your fleet is static and you need fast deployments. Not a best practice and not terribly secure, but you mentioned similar scheme in your post
Using AWS Image Builder (https://aws.amazon.com/image-builder/) or Packer (https://www.packer.io/) to build new release image and then replace old image with new in your EC2 autoscaling group
Build docker image of your app and use ECS or EKS to host it. I would recommend this approach if you're not married to running code directly on EC2 hosts.
There's a similar question here but from 2018 were the solution requires changing the base image for the workers. Another suggestion is to ssh into each node and apt-get install there. This doesn't seem useful because when auto scale spawns new nodes, you'd need to do it again and again.
Anyway, is there a reasonable way to upgrade the base gcloud in late 2020?
Because task instances run in a shared execution environment, it is generally not recommended to use the gcloud CLI within Composer Airflow tasks, when possible, to avoid state or version conflicts. For example, if you have multiple users using the same Cloud Composer environment, and either of them changes the active credentials used by gcloud, then they can unknowingly break the other's workflows.
Instead, consider using the Cloud SDK Python libraries to do what you need to do programmatically, or use the airflow.providers.google.cloud operators, which may already have what you need.
If you really need to use the gcloud CLI and don't share the environment, then you can use a BashOperator with a install/upgrade script to create a prerequisite for any tasks that need to use the CLI. Alternatively, you can build a custom Docker image with gcloud installed, and use GKEPodOperator or KubernetesPodOperator to run a Kubernetes pod to run the CLI command. That would be slower, but more reliable than verifying dependencies each time.
I have a Python AWS Lambda running on a Linux, but due to some dependencies, I need it to be deployed on a Windows. I have tried using Python Azure Functions and have successfully deployed it on a Linux as well, but found out they cannot be deployed on Windows. Is it possible to do it with AWS Lambda?
Basically my solution has a few .exe that need to be run by a python library (Tesseract OCR and pytesseract)
AWS Lambda and Azure Functions are considered Function as a Service (FaaS) solutions, where the developer worries about the code and the cloud provider worries about availability, scalability and the platform underneath to run the code.
In that aspect, you can't run any of them on a server. If you need specific Windows dependencies, you must create a Python project as you normally would, install the dependencies and configure the Windows Server, being responsible for infrastructure and OS configurations and management.
I have a python application that I've created an executable of, using pyinstaller. The entire python interpreter is packaged into the executable with all its pip dependencies.
So now my application can run in environments where python or python modules may not be installed, but there are still some dependencies:
1) MongoDB - This is the database my application uses, and it needs to be installed on a system for it to work of course.
2) Mosquitto - This service is required because the application uses MQTT to receive/send commands.
My current method of handling this is to use a shell script which installs mongodb and mosquitto the first time when my application is deployed somewhere. I just discovered docker, and I was wondering if it is capable of packaging these 'external' dependencies into a docker image?
Is it possible for me to have one standalone "thing" which will run in any environment regardless of whether mongoDB or mosquitto are installed there?
And how exactly would I go about doing this?
(Unrelated but this application is meant to run on a raspberry pi)
If you adopted Docker here:
You'd still have to "separately" run the external services; they couldn't be packaged into a single artifact per se. There's a standard tool called Docker Compose that provides this capability, though, and you'd generally distribute a docker-compose.yml file that describes how to run the set of related containers.
It's unusual to distribute a Docker image as files; instead you'd push your built image to a registry (like Docker Hub, but the major public-cloud providers offer this as a hosted service, there are a couple of independent services, or you can run your own). Docker can then retrieve the image via HTTP.
Docker containers can only be run by root-equivalent users. Since you're talking about installing databases as part of your bringup process this probably isn't a concern for you, but you could run a plain-Python or pyinstallered application as an ordinary user. Anyone who can run any Docker command has unrestricted root-level access on the host.
So I was looking into scheduling a python script on a daily basis and, rather than using Task Scheduler on my own machine, I was wondering if it is possible to do so using an Azure cloud account.
For your needs, I suggest you use Web Jobs in Web Apps Service.
It has two types of Azure Web Jobs for you to choose:
Continuous and Trigger.
For your needs, Trigger should be adopted.
You could refer to the document here for more details.In addition, here shows how to run tasks in WebJobs.
You could refer to the steps as below to create your webjob.
Step 1 :
Use the virtualenv component to create an independent python runtime environment in your system.Please install it first with command pip install virtualenv if you don't have it.
If you installed it successfully ,you could see it in your python/Scripts file.
Step2 : Run the commad to create independent python runtime environment.
Step 3: Then go into the created directory's Scripts folder and activate it (this step is important , don't miss it)
Please don't close this command window and use pip install <your libraryname> to download external libraries in this command window.
Step 4：Keep the Webjob.py(which is your own business code) uniformly compressed into a folder with the libs packages in the Libs/site-packages folder that you rely on.
Step 5:
Create webjob in Web app service and upload the zip file,then you could execute your Web Job and check the log
You could also refer to the SO thread:
1.Options for running Python scripts in Azure
2.Python libraries on Web Job
BTW,you need to create a azure web app first because Webjob runs in azure web app.
Hope it helps you.