See buildInputs value in the file default.nix to get the list of requiered tools, also review the environment variables defined in that file.

After building alephium-blockflow you can start it as the following example:

ALEPHIUM_ENV=prod groups=4 brokerNum=2 nodes=4 ./make run

One should first create an AMI (Amazon Machine Image) for a given package version, this image will then be used to create one or many VPC clusters.

• image-create
• cluster-create
• cluster-delete
• image-delete

Note: every resources created by image-create are tagged using $PACKAGE_NAME-$PACKAGE_VERSION, similarly all resources created by cluster-create are tagged using $CLUSTER_ID.

• Populate .aws/config (copy it from .aws/config.sample)

  • https://www.cloudberrylab.com/blog/how-to-find-your-aws-access-key-id-and-secret-access-key-and-register-with-cloudberry-s3-explorer/

• Add default.pem

  • https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/ec2-key-pairs.html#having-ec2-create-your-key-pair

• Copy settings.json.tmpl as settings.json and populate accordingly

  • Add package file <packageId>.zip in project root

• Build the AMI image with ./image-create

• Once the image is ready, one can start a new cluster with ./cluster-create.sh $CLUSTER_ID

The EC2 console ("Instances") show public IP addresses of nodes, which can be used to communicate with them.

One can SSH to them, and print the log output.

ssh -i default.pem ec2-user@$IP_ADDRESS tail -f /var/log/cloud-init-output.log

One can also use pssh to access multiple instances at once.

./cluster-pssh.sh foo "tail /root/.alephium/logs/alephium.log"

One can do RPC call accross the whole cluster, for example to start mining.

./cluster-rpc-call foo mining/start

You can attach a JVisualVM session to any node, first open a ssh tunnel with the following port bindings.

ssh -L 9010:localhost:9010 -L 9011:localhost:9011 -i default.pem ec2-user@

You can now open a remote JMX connection is JVisualVM using localhost:9010.

• Upload the application to a S3 bucket, and then fetch from there.
  • Add cluster-update

You can delete the cluster once you are done with it using ./cluster-delete $CLUSTER_ID.

In order to completely remove an image from the ELS storage, one can run ./image-delete.

This can be used to recreate the image of a same package version with an update package file.