Install Agent

We assume that your system is configured with a sudoable admin user named devops. Your Backend.AI manager should be already set up and running.

Guide variables

⚠️ Prepare the values of the following variables before working with this page and replace their occurrences with the values when you follow the guide.

{NS}	The etcd namespace (just create a unique string like domain names)
{ETCDADDR}	The etcd cluster address ({ETCDHOST}:{ETCDPORT}, localhost:2379 for development setup)

Optional variables

{SSLCERT}	The path to your SSL certificate (bundled with CA chain certificates)
{SSLPKEY}	The path to your SSL private key
{S3AKEY}	The access key for AWS S3 or compatible services[1]
{S3SKEY}	The secret key for AWS S3 or compatible services
{DDAPIKEY}	The Datadog API key
{DDAPPKEY}	The Datadog application key
{SENTRYURL}	The private Sentry report URL

Install dependencies for daemonization

Ubuntu

$ sudo apt-get -y update
$ sudo apt-get -y dist-upgrade
$ sudo apt-get install -y ca-certificates git-core supervisor

Here are some optional but useful packages:

$ sudo apt-get install -y vim tmux htop

CentOS / RHEL

(TODO)

Prepare CUDA (if available)

Check out the Install CUDA guide.

Prepare Python 3.6+

Check out Install Python via pyenv for instructions. Create a virtualenv named venv-agent.

Install Backend.AI Agent as Package

$ pyenv shell venv-agent
$ pip install -U setuptools pip
$ pip install -U backend.ai-agent

Monitoring and Logging

Check out the Install Monitoring and Logging Tools guide.

Configure supervisord

supervisord application config

$ sudo vi /etc/supervisor/conf.d/apps.conf

[program:backendai-agent]
user = devops
stopsignal = TERM
stopasgroup = true
command = /home/devops/run-agent.sh

pyenv + venv initialization script for non-login shells

$ vi /home/devops/init-venv.sh

#!/bin/bash
export PYENV_ROOT="$HOME/.pyenv"
export PATH="$PYENV_ROOT/bin:$PATH"
eval "$(pyenv init -)"
eval "$(pyenv virtualenv-init -)"
pyenv shell venv-agent

Prepare scratch directory (place for kernel containers' /home/work)

$ sudo mkdir -p /var/cache/scratches
$ sudo chown devops:devops /var/cache/scratches

The main program managed by supervisord

$ vi /home/devops/run-agent.sh

source /home/devops/init-venv.sh
umask 0002
export AWS_ACCESS_KEY_ID="{S3AKEY}"
export AWS_SECRET_ACCESS_KEY="{S3SEKEY}"
export DATADOG_API_KEY={DDAPIKEY}
export DATADOG_APP_KEY={DDAPPKEY}
export RAVEN_URI="{SENTRYURL}"
exec python -m ai.backend.agent.server \
            --etcd-addr {ETCDADDR} \
            --namespace {NS} \
            --scratch-root=/var/cache/scratches

Prepare Kernel Images

You need to pull the kernel container images first to actually spawn compute sessions.

$ docker pull lablup/kernel-python:latest
$ docker pull lablup/kernel-python:3.6

For the full list of publicly available kernels, check out the kernels repository.

Finally, Run!

$ sudo supervisorctl reread
$ sudo supervisorctl start backendai-agent