Wercker on CoreOS

Transcript

1. wercker
2. wercker on CoreOS http://wercker.com @wercker
3. Wercker: automation driven development • Wercker is an automation platform • ci/cd not ambitious enough • continuous is assumed
4. Interface • wercker.yml • Describe job environment • Describe a work pipeline • Describe required services
5. Interface • wercker.yml • Describe job environment • Describe a work pipeline • Describe required services box:  tcnksm/gox@0.1.0
6. Interface • wercker.yml • Describe job environment • Describe a work pipeline • Describe required services steps:              -­‐  script:                      name:  install                      code:  go  get  -­‐v  -­‐d  ./...              -­‐  golint
7. Interface • wercker.yml • Describe job environment • Describe a work pipeline • Describe required services services:          -­‐  mongodb
8. Output • Pipeline outputs container image • Hugely beneficial • Immutable, tested object linked to commit • Repeatability
9. Under the hood • Standalone process dispatches work to fleet • Called “kiddie-pool”
10. mhook • Needed simple package manager • downloads packages from S3 • indexed by project, branch, and commit • similar to Kelsey Hightower’s `pm` https://github.com/wercker/mhook
11. kiddie-pool uses mhook [Unit]   Description=Wercker  Kiddie-­‐Pool  %i   After=install-­‐mhook.service   [Service]   ExecStartPre=/tmp/mhook  -­‐-­‐bucket  <bucket>  \\          —project  kiddie-­‐pool  \\   —branch  master  \\   —commit  latest  \\   /opt/bin/kiddie-­‐pool-­‐%i   ExecStartPre=/usr/bin/chmod  +x  /tmp/kiddie-­‐pool-­‐%i  ...   [Install]   WantedBy=multi-­‐user.target   [X-­‐Fleet]   Global=true
12. kiddie-pool • Sits between work queue and fleet • Constructs Unit from job metadata • Unit invokes `wercker` cli tool • Sends to fleet over http API
13. wercker cli • Understands how to actually run the job • You can download and run wercker locally • rapid iteration • don’t even need us • wow http://wercker.com/downloads/
14. What’s next • Smart job assignment • Fleet naively allocates work • Need a job scheduler that’s smarter • Colocate jobs, data, and services • Help us build this
15. What’s next • mhook • Push behavior • Signing packages • Wercker step
16. Logging • Logging • Use Graylog because they have a working docker image • `journalhook` sends application data to journal • `journal2gelf` container streams journal to graylog • persists journal cursor https://github.com/wercker/journalhook
17. Provisioning • Vagrant for development • AWS/Cloud Formation for user-facing • Single script to send application units to fleet • Separation of concerns, no duplication
18. etcd • Used to publish service information • In the application to signal job aborts • Best way to run is as cluster of its own • should get easier with 2.0 with proxies • need to configure many things to have correct etcd url
19. References 1. http://wercker.com/downloads/ 2. https://github.com/wercker/mhook 3. https://github.com/wercker/journalhook 4. https://github.com/systemd/journal2gelf 5. https://registry.hub.docker.com/u/mwhooker/ journal2gelf/ 6. https://speakerdeck.com/mwhooker/wercker-on- coreos
20. Thank you Matt Hooker wercker Software Engineer matt@wercker.com @mwhooker http://wercker.com http://twitter.com/wercker Grab a shirt and/or sticker! wercker