distributivenetwork/kube-evaluator repository overview

⁠Kube-Evaluator Container

This container is meant to be used for kubernetes style evaluator/supervisor set ups. In these style of set ups, the evaluator is separate from the evaluator and thus this docker image does not immediately start up an evaluator. It is left up to the manifest to start the evaluator or the supervisor.

docker pull distributivenetwork/kube-evaluator:latest

--------8<--------

To launch the evaluator using this docker image, your entrypoint will have to launch ~/bin/dcp-evaluator-start -e ~/evaluator-bin/dcp-evaluator .....

To launch the worker using this docker image, your entrypoint will have to launch ~/bin/dcp-worker .....

-------->8--------

The default keystore is used to identify the DCP Bank Account into which compute credits that are earned by the worker are deposited. If you're spending compute credits to get a job done you'll need to use the default keystore to authorize the withdrawal from your account as it will have the private key that allows access to it, however, if you're only doing work and earning credits, you only need the public key associated with the default keystore, which uniquely identifies the account on the bank. You must create a Bank Account on the portal (found at https://dcp.cloud⁠) then copy its public address into your clipboard and paste it into the command line you use to start the docker worker. As the worker does work and earns compute credits, you'll be able to monitor your bank balance in the portal. You can in fact pass any of the arguments above on the docker command line and they'll be read by the worker which will behave accordingly.

An example command doing this would be:

docker run distributivenetwork/kube-evaluator:latest ~/bin/dcp-worker --paymentAddress=<bankID>

-------->8--------

If you're not familiar with how to launch a worker or an evaluator, this is not the image for you. Instead please take a look at this docker image⁠.