Backend.AI Agent
Project description
Backend.AI Agent
The Backend.AI Agent is a small daemon that does:
- Reports the status and available resource slots of a worker to the manager
- Routes code execution requests to the designated kernel container
- Manages the lifecycle of kernel containers (create/monitor/destroy them)
Package Structure
ai.backendagent: The agent daemon implementation
Installation
Please visit the installation guides.
For development
Prerequisites
libnsappy-devorsnappy-develsystem package depending on your distro- Python 3.6 or higher with pyenv and pyenv-virtualenv (optional but recommneded)
- Docker 17.03 or later with docker-compose (18.03 or later is recommended)
Clone the meta repository and install a "halfstack" configuration. The halfstack configuration installs and runs dependency daemons such as etcd in the background.
~$ git clone https://github.com/lablup/backend.ai halfstack
~$ cd halfstack
~/halfstack$ docker-compose -f docker-compose.halfstack.yml up -d
Then prepare the source clone of the agent as follows. First install the current working copy.
~$ git clone https://github.com/lablup/backend.ai-agent agent
~$ cd agent
~/agent$ pyenv virtualenv venv-agent
~/agent$ pyenv local venv-agent
~/agent (venv-agent) $ pip install -U pip setuptools # ensure latest versions
~/agent (venv-agent) $ pip install -U -r requirements-dev.txt
Before running, you first need to prepare "the kernel runner environment", which is composed of a dedicated Docker image that is mounted into kernel containers at runtime. Since our kernel images have two different base Linux distros, Alpine and Ubuntu, you need to build/download the krunner-env images twice as follows.
For development:
~/agent (venv-agent) $ python -m ai.backend.agent.kernel build-krunner-env alpine3.8
~/agent (venv-agent) $ python -m ai.backend.agent.kernel build-krunner-env ubuntu16.04
or you pull the matching version from the Docker Hub (only supported for already released versions):
~/agent (venv-agent) $ docker pull lablup/backendai-krunner-env:19.03-alpine3.8
~/agent (venv-agent) $ docker pull lablup/backendai-krunner-env:19.03-ubuntu16.04
With the halfstack, you can run the agent simply like (note that you need a working manager running with the halfstack already):
~/agent (venv-agent) $ scripts/run-with-halfstack.sh python -m ai.backend.agent.server \
> --scratch-root=/tmp --debug
To run tests:
~/agent (venv-agent) $ scripts/run-with-halfstack.sh python -m pytest -m 'not integration'
To run tests including integration tests, you first need to install the manager in the same virtualenv.
~$ git clone https://github.com/lablup/backend.ai-manager manager
~$ cd agent
~/agent (venv-agent) $ pip install -e ../manager
~/agent (venv-agent) $ scripts/run-with-halfstack.sh python -m pytest
Deployment
Running from a command line
The minimal command to execute:
python -m ai.backend.agent.server --etcd-addr localhost:2379 --namespace my-cluster
The agent reads most configurations from the given etcd v3 server where the cluster administrator or the Backend.AI manager stores all the necessary settings.
The etcd address and namespace must match with the manager to make the agent paired and activated. By specifying distinguished namespaces, you may share a single etcd cluster with multiple separate Backend.AI clusters.
By default the agent uses /var/cache/scratches directory for making temporary
home directories used by kernel containers (the /home/work volume mounted in
containers). Note that the directory must exist in prior and the agent-running
user must have ownership of it. You can change the location by
--scratch-root option.
For more arguments and options, run the command with --help option.
Example config for agent server/instances
/etc/supervisor/conf.d/agent.conf:
[program:backend.ai-agent]
user = user
stopsignal = TERM
stopasgroup = true
command = /home/user/run-agent.sh
/home/user/run-agent.sh:
#!/bin/sh
source /home/user/venv/bin/activate
exec python -m ai.backend.agent.server \
--etcd-addr localhost:2379 \
--namespace my-cluster
Networking
Basically all TCP ports must be transparently open to the manager. The manager and agent should run in the same local network or different networks reachable via VPNs.
The operation of agent itself does not require both incoming/outgoing access to the public Internet, but if the user's computation programs need, the docker containers should be able to access the public Internet (maybe via some corporate firewalls).
Several optional features such as automatic kernel image updates may require outgoing public Internet access from the agent as well.
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
File details
Details for the file backend.ai-agent-19.3.0b3.tar.gz.
File metadata
- Download URL: backend.ai-agent-19.3.0b3.tar.gz
- Upload date:
- Size: 6.0 MB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/1.12.1 pkginfo/1.4.2 requests/2.21.0 setuptools/40.7.1 requests-toolbelt/0.8.0 tqdm/4.28.1 CPython/3.6.7
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 | b7f8fd9a48b3ea024c35ae618e08ce08fd5d17932256e9976b4ef24e88f63826 |
|
| MD5 | 16d474d85837c7563d9174725c2ca34f |
|
| BLAKE2b-256 | 3b074434695ee92a3d711645d6e28f40a87affc5a84d412f6bf897b7b41c4bbc |
Provenance
File details
Details for the file backend.ai_agent-19.3.0b3-py3-none-any.whl.
File metadata
- Download URL: backend.ai_agent-19.3.0b3-py3-none-any.whl
- Upload date:
- Size: 6.1 MB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/1.12.1 pkginfo/1.4.2 requests/2.21.0 setuptools/40.7.1 requests-toolbelt/0.8.0 tqdm/4.28.1 CPython/3.6.7
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 | afb26316efd3c6a61e8f445520ab31c80c5e7879f03ab848ec487473be96edeb |
|
| MD5 | f4edcec72aefcc4928df176a9ce284bc |
|
| BLAKE2b-256 | fff9b63437959837b65e80a706a52a2b42ab31fa9b127cbc98db96058f13effb |
