Complementary Automated Methods for Proofreading
- Set up AWS credentials for IARPA MICrONS EXT AWS account
- Download
kubectl
,eksctl
, andawscli
- Update kubeconfig
aws eks --region region update-kubeconfig --name agents-cluster
- Copy
~/.aws/credentials
,~/.cloudvolume/secrets/cave-secret.json
, and~/.neuvuequeue/neuvuequeue.cfg
to/secrets
. - Build app docker image and push to ECR.
- Populate queue with the channels you want to transfer.
- Create a nodegroup using
eksctl
to add compute to your cluster.
eksctl create nodegroup --config-file=kube/{cpu|gpu}-nodegroup.yml
- If the app/ code changed, update the docker image on AWS ECR.
- Verify that
kubectl
can connect to your cluster withkubectl get nodes
andkubectl get pods
. - Run the job with
kubectl apply -f kube/{agents|gpu}-deploy.yaml
- To enter an interactive terminal for a pod
kubectl exec -it [pod-name] -- /bin/bash/
- Delete a node group
eksctl delete nodegroup --region=us-east-1 --cluster="spdb-to-cv-2" --name=spdb-to-cv-workers-2x
- Delete a job
kubectl delete jobs tips-jobs
- Add user to masters list
eksctl create iamidentitymapping --cluster spdb-to-cv-2 --arn arn:aws:iam::407510763690:user/hiderrt1-bossdb --username hiderrt1-bossdb --group system:masters --region us-east-1
- Get masters list
eksctl get iamidentitymapping --cluster spdb-to-cv-2 --region us-east-1
- Get CPU/MEM usage
kubectl top nodes --tail
- Get logs for a pod
kubectl logs {pod_name}