Skip to content

Latest commit

 

History

History
80 lines (51 loc) · 3.25 KB

AWS_README.md

File metadata and controls

80 lines (51 loc) · 3.25 KB
Raw

1. Create an AWS Account

https://portal.aws.amazon.com/billing/signup#/start

NOTE: You will need to use a credit cart to set up an account (even with free resources).

2. Create an Access Key:

Part way through the tutorial you will need an "AWS Access Key ID" and "AWS Secret Access Key".

Open AWS console in a browser. In the top right of the screen click your user name and then "My Security Credentials". On the security credentials page, click "Access Keys". Then click create a new access key. Select Download Access key. Then open the file so that when the time comes you can put use the access key in the tutorial.

3. Find your Region

If you are fine hosting your cluster in your default region you can proceed to the next step. Otherwise find your region code here:

https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/using-regions-availability-zones.html#concepts-available-regions

Dirk used east-us-2 and Kai used us-west-2

Install Software

Follow the tutorial until you get to the part that tells you to run kubectl get svc. Run that command and ensure everything is working as expected. Note that in the tutorial you will need to use some of the information you obtained in the previous parts of this document. Furthermore, when the time comes select "AWS for linux-only workloads". Basic Steps include:

  • Download/install AWS CLI
  • Download/install the eksctl program
  • Download/install the AWS Kuberneties installer

Note that running the command that creates a kubernetes cluster typically takes about 15 minutes.

https://docs.aws.amazon.com/eks/latest/userguide/getting-started-eksctl.html

Connect your computer to AWS

Now run the following CLI command to set up the AWS account on your local machine. You will need your Access key ID, secret Access Key and your selected region (use json for format) aws configure

Start a cluster

Use the following command to start a cluster. (You may want to remove the ssh lines)

eksctl create cluster \
--name prod \
--version 1.16 \
--region us-east-2 \
--nodegroup-name standard-workers \
--node-type t3.medium \
--nodes 3 \
--nodes-min 1 \
--nodes-max 4 \
--ssh-access \
--ssh-public-key ~/.ssh/id_rsa.pub \
--managed

Running See Segment

Once this has successfully finished you can start the application using kubectl

Navigate to ./image-segmenter-finder/src/see_server/kube_commands in your cloned local version of this repository.

Then run:

kubectl apply -f server_service.yaml

kubectl apply -f server.yaml

`kubectl apply -f segmentation_job.yaml.``

Everything should now be up and running. You can access the server by typing:

kubectl get services

It takes a minute to get set up but eventually there will be an external ip for the server_service By copying this ip into a browser you will be able to access the server.

You should be able to delete your cluster with a command similar to

eksctl delete cluster --region=us-west-2 --name=prod

Known Issues:

When creating the kubernetes cluster it may state that the creation of the cluster failed because your account is not authorized to use the required compute resources. In this case retrying the command ended up fixing the issue.