Building and Deploying MediaSalsa an Open source Digital Asset Management platform as a Service. My FlossUK 2014 talk

1. Building and Deploying MediaSalsa
A Digital Asset Management System as a Service
Kris Buytaert
@krisbuytaert
Slides by Michel van de Ven and
Julien Pivotto

2. Kris BuytaertKris Buytaert
● I used to be a Dev,I used to be a Dev,
● Then Became an OpThen Became an Op
● Chief Trolling Officer and Open SourceChief Trolling Officer and Open Source
Consultant @inuits.euConsultant @inuits.eu
● Everything is an effing DNS ProblemEverything is an effing DNS Problem
● Building Clouds since before the bookstoreBuilding Clouds since before the bookstore
● Some books, some papers, some blogsSome books, some papers, some blogs
● Evangelizing devopsEvangelizing devops

3. MediaMosa
● Drupal-based Digital Asset Management system
● Commissioned by SURFnet and Kennisnet
● Open Source (GPLv2), Open Standards
● Webservice oriented: REST
● Store & retrieve assets
● Manage metadata using open standards
– Dublin Core, Qualified DC, IEEE/LOM, CZP
– OAI-PMH, BagIt
● Transcode video, audio, images, PDF &c, stream content
● Users: Kennisnet, NIBG, UGent, UvA, TiU, RUG, Radboud, UOslo,
Avans, PolitieAcademie, Acquia (NBC Sport), Cineca &c
● http://mediamosa.org

5. =
MediaMosa as a Service

6. MediaSalsa infrastructure (simplified)
● For each environment (OTAP)
– Backends: Core service (MediaMosa)
– Frontends: Optional
– Web servers
– Database servers
– Solr servers
– Transcoding servers
● Instance @ RUG DC for NL .edu
● Instances @ Commercial DCs for non-edu

7. Culture
Automation
Measurement
Sharing

8. Inuits
● Inuits is an Open Source company
– We contribute back
● +40 people in 3 countries (BE, NL, UA)
● One language: English
● We offer
– Consulting
● Development
● System Administration
– and a niche Saas Platform

9. Ideal world vs budget and reality
→ pragmatic approach

10. Distributed team
Communication = HELL
How to fix it?

11. Distributed team
● Daily virtual stand-up over XMPP / Hangout
● Redmine project management
– Issues, tasks, sprints
– Repositories
– Documentation (wiki)
● Internal mailing lists
● Jenkins notifications by mail and XMPP
● Internal training

12. Dev and Ops
● Dev
● Architecture
● Develop new features
● Write unit tests (!)
● Ops
● Infrastructure (puppet)
● CD and CI (jenkins)
→ Ops teach dev about monitoring and
distributed services
→ Dev teach ops about required libraries and
testing, dependencies

13. Culture
Automation
Measurement
Sharing

14. Puppet
Puppet automates all the things
→ mcollective orchestrates all the things

15. CD
● Continuous Delivery vs Continuous Deployment
– “Continuous Delivery doesn't mean every change is deployed to production
ASAP. It means every change is proven to be deployable at any time”
(@ccaum)
● Puppet code
– Deployed to dev environment
– Same puppet code for each environment
– User-triggered deployments to UAT & Prod
– Feature flags in Puppet code per environment (switchable architecture)
● Application code
– Continuous integration in dev
– User-triggered deployments to UAT
– Deployment to prod is a business decision

16. Testing
● Developers test a lot, but
– The tests don’t work
– It works on my machine™
– Wrong platform
– Wrong PHP version
Fixed now, thanks to Jenkins!
and Vagrant...

17. Version Control
● Git
● Code is under revision control
– Prefer small commits
– Local features branches
● Infrastructure as code → git / hiera

18. Using OS packaging system
● Consistency, security, dependencies
● Uniquely identify where files are coming from
● Source repo may not be reacheable
● Little overhead when you automate
● Configuration does not belong in a package

19. Pipelines
● A collection of jobs
● Run in sequence
● Start on checkout, end on deployment
● From the developers’ side:
→ Git push
← Mail with changes + link to deploy

20. Our current pipelines
● Puppetized
● Puppet code and Application code (backend and
frontends)
● 1 pipeline / project / customer
Defining a pipeline:
salsajobs::pipeline::frontend {
’qwerty-inc’:
git_repository => ’ssh://git@redmine/qwerty.git’,
dashboard_view => ’mds-frontends’,
target_urls => {
’uat’ => ’http://qwerty.uat.mediasalsa.eu’,
’prod’ => ’http://qwerty.mediasalsa.eu’,
},
require => Salsajobs::Views::Dashboard[’mds-frontends’],
}

21. Pipelines steps
● Checkout
● Syntax: php -l
● Style: Drupal Coder
● Package: FPM
● Deploy to dev environment: mcollective
● Tests in dev environment: drush run-tests
● Publish package and promote: mcollective

22. Promotion
● At the end of the pipeline
● Send a simple email
● A link to the promotion page
● The changelog
● Promotion page contains one button per environment
● Must promote to UAT before Production

23. Tools used with Jenkins
● Pulp to manage RPM repositories
● Mcollective to update packages, run drush

24. Same Pipelines, Tools, Patterns
are used by both devs and ops

25. Culture
Automation
Measurement
Sharing

26. Logstash
● Collect all the logs
– Drupal logs
– Apache logs
– Deployment logs
– System logs
● Interprete, filter and correlate them
● Logstash, ElasticSearch, Kibana, statsd, Graphite

27. Icinga
● Monitor everything
– vhosts
– databases
– cronjobs
– unit test suites

28. Graphite + gdash
● Collectd
● Monitor platform usage
● FFmpeg usage
● Number of accounts
● Pipelined !

29. Culture
Automation
Measurement
Sharing

30. Open Source
● Mediamosa is Fully Open Source
● Lots of the PuppetCode to deploy it is
● Our passwords etc aren't

31. Experiences
We have just shared our experience...
But also at the NLUUG Fall Conference, DrupalCon Prague

32. ContactContact
Kris BuytaertKris Buytaert Kris.Buytaert@inuits.beKris.Buytaert@inuits.be
Further ReadingFurther Reading
@krisbuytaert@krisbuytaert
http://www.krisbuytaert.be/blog/http://www.krisbuytaert.be/blog/
http://www.inuits.be/http://www.inuits.be/
Inuits HQInuits HQ
Duboistraat 50Duboistraat 50
2060 Antwerpen2060 Antwerpen
BelgiumBelgium
891.514.231891.514.231
+32 475 961221+32 475 961221