https://github.com/milosgajdos/taurus
Experimental Mesos framework for Docker containers written in Go
https://github.com/milosgajdos/taurus
docker-container go golang mesos mesos-framework
Last synced: about 1 month ago
JSON representation
Experimental Mesos framework for Docker containers written in Go
- Host: GitHub
- URL: https://github.com/milosgajdos/taurus
- Owner: milosgajdos
- Created: 2015-05-24T01:35:37.000Z (about 10 years ago)
- Default Branch: master
- Last Pushed: 2020-02-25T21:03:09.000Z (about 5 years ago)
- Last Synced: 2024-12-26T05:42:09.590Z (5 months ago)
- Topics: docker-container, go, golang, mesos, mesos-framework
- Language: Go
- Homepage:
- Size: 303 KB
- Stars: 15
- Watchers: 2
- Forks: 5
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
**THIS PROJECT IS MASSIVELY ALPHA AND UNSTABLE AT THE MOMENT. CONSIDER YOURSELF WARNED.**
# taurus
Simple Mesos Container framework written in Go.
At the moment this is just a **PoC** serving for learning purpose. I'm not sure if I turn into more than just that.
There is a lot of design flaws - lots of existing code would need to be refactored to make this useful.
## Theory of operation
**taurus** provides REST API that allows you to submit taurus Jobs and run them as Mesos tasks in Docker containers.
You can run **taurus** using docker-compose.
Out of the box:
- **taurus** uses [NATS](https://github.com/nats-io/nats) distributed queue to queue pending tasks. You can implement your own queue that satisfies taurus ```queue``` interface and replace the basic queue implementation.
- **taurus** stores Jobs in a local store implemented using [gkvlite](https://github.com/steveyen/gkvlite). You can implement your own Job store that satisfies taurus ```store``` Go interface and replace the basic store implementation.
- **taurus** implements a simple scheduler worker which does all the heavy lifting i.e.: detecting state of submitted Jobs and taking appropriate action like queueing them to particular queues, killing the stopped Job tasks etc. You can implement your own worker which satisfies taurus ```worker``` interface and replace the basic worker implementation.
## Example
There are some example jobs in ```examplejobs``` directory.
Submit a job:
```
$ curl -i -X POST 127.0.0.1:8080/job/new -d @job.json
HTTP/1.1 200 OK
Content-Type: application/json; charset=UTF-8
Date: Sun, 12 Jul 2015 22:53:17 GMT
Content-Length: 0
```
Taurus logs:
```
2015/07/12 23:53:17 server.go:43: POST /job/new
2015/07/12 23:53:17 api.go:103: Submitting Job taurusjob
2015/07/12 23:53:17 scheduler.go:315: No tasks to kill
2015/07/12 23:53:17 scheduler.go:381: Launched tasks: map[string]string{}
2015/07/12 23:53:17 scheduler.go:568: No Pending tasks available
2015/07/12 23:53:18 scheduler.go:159: Creating new Pending task taurusjob-web-prod-1436741598-c9cbf28d-28e8-11e5-a327-600308a80bb2 for job taurusjob
2015/07/12 23:53:18 scheduler.go:159: Creating new Pending task taurusjob-web-prod-1436741598-c9cc085d-28e8-11e5-a327-600308a80bb2 for job taurusjob
2015/07/12 23:53:18 scheduler.go:159: Creating new Pending task taurusjob-web-prod-1436741598-c9cc0d1c-28e8-11e5-a327-600308a80bb2 for job taurusjob
2015/07/12 23:53:18 scheduler.go:159: Creating new Pending task taurusjob-web-prod-1436741598-c9cc10ea-28e8-11e5-a327-600308a80bb2 for job taurusjob
2015/07/12 23:53:18 scheduler.go:203: Queueing task taurusjob-web-prod-1436741598-c9cbf28d-28e8-11e5-a327-600308a80bb2 to Pending queue
2015/07/12 23:53:18 scheduler.go:203: Queueing task taurusjob-web-prod-1436741598-c9cc085d-28e8-11e5-a327-600308a80bb2 to Pending queue
2015/07/12 23:53:18 scheduler.go:203: Queueing task taurusjob-web-prod-1436741598-c9cc0d1c-28e8-11e5-a327-600308a80bb2 to Pending queue
2015/07/12 23:53:18 scheduler.go:203: Queueing task taurusjob-web-prod-1436741598-c9cc10ea-28e8-11e5-a327-600308a80bb2 to Pending queue
...
...
2015/07/12 23:53:26 scheduler.go:667: Task taurusjob-web-prod-1436741598-c9cc0d1c-28e8-11e5-a327-600308a80bb2 is in state TASK_RUNNING
2015/07/12 23:53:26 scheduler.go:667: Task taurusjob-web-prod-1436741598-c9cc085d-28e8-11e5-a327-600308a80bb2 is in state TASK_RUNNING
2015/07/12 23:53:26 scheduler.go:667: Task taurusjob-web-prod-1436741598-c9cbf28d-28e8-11e5-a327-600308a80bb2 is in state TASK_RUNNING
2015/07/12 23:53:26 scheduler.go:667: Task taurusjob-web-prod-1436741598-c9cc10ea-28e8-11e5-a327-600308a80bb2 is in state TASK_RUNNING
```
Check if the containers are running:
```
$ docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
07bf0188afc6 nginx:latest "nginx -g 'daemon of 8 seconds ago Up 6 seconds 443/tcp, 80/tcp mesos-466f9aa8-32e1-4148-9e42-c0716f8c594a
4e246cfab0f7 nginx:latest "nginx -g 'daemon of 8 seconds ago Up 6 seconds 443/tcp, 80/tcp mesos-78df44ea-2ca6-448c-9dbf-9cf88081b8fc
af6fefad032c nginx:latest "nginx -g 'daemon of 8 seconds ago Up 6 seconds 443/tcp, 80/tcp mesos-a689eec3-6d63-44fb-818a-64c427b2b88a
98b0902a8aae nginx:latest "nginx -g 'daemon of 9 seconds ago Up 8 seconds 443/tcp, 80/tcp mesos-f8534b44-4c88-43b9-aced-21bee4c91f8a
```
Stop the job:
```
$ curl -i -X DELETE 127.0.0.1:8080/job/taurusjob
HTTP/1.1 200 OK
Content-Type: application/json; charset=UTF-8
Date: Sun, 12 Jul 2015 22:54:39 GMT
Content-Length: 0
```
Taurus logs:
```
2015/07/12 23:54:39 server.go:43: DELETE /job/taurusjob
2015/07/12 23:54:39 api.go:131: Stopping Job taurusjob
2015/07/12 23:54:39 api.go:148: Killing Job taurusjob
2015/07/12 23:54:39 scheduler.go:258: Queueing task taurusjob-web-prod-1436741598-c9cc10ea-28e8-11e5-a327-600308a80bb2 to Doomed queue
2015/07/12 23:54:39 scheduler.go:258: Queueing task taurusjob-web-prod-1436741598-c9cc0d1c-28e8-11e5-a327-600308a80bb2 to Doomed queue
2015/07/12 23:54:39 scheduler.go:258: Queueing task taurusjob-web-prod-1436741598-c9cc085d-28e8-11e5-a327-600308a80bb2 to Doomed queue
2015/07/12 23:54:39 scheduler.go:258: Queueing task taurusjob-web-prod-1436741598-c9cbf28d-28e8-11e5-a327-600308a80bb2 to Doomed queue
...
...
2015/07/12 23:54:42 scheduler.go:667: Task taurusjob-web-prod-1436741598-c9cc0d1c-28e8-11e5-a327-600308a80bb2 is in state TASK_KILLED
2015/07/12 23:54:42 scheduler.go:667: Task taurusjob-web-prod-1436741598-c9cbf28d-28e8-11e5-a327-600308a80bb2 is in state TASK_KILLED
2015/07/12 23:54:42 scheduler.go:667: Task taurusjob-web-prod-1436741598-c9cc10ea-28e8-11e5-a327-600308a80bb2 is in state TASK_KILLED
2015/07/12 23:54:42 scheduler.go:667: Task taurusjob-web-prod-1436741598-c9cc085d-28e8-11e5-a327-600308a80bb2 is in state TASK_KILLED
```
Check if the containers are running:
```
$ docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
```
# TODO
**E_TOO_MUCH**, but in a gist:
- redesign a lot of things (I might spin another project from this)
- implement etcd/consul/zk storage driver support
- implement more scalable solution using something like graft by Apcera
AND THE LIST GOES ON AND ON AND ON