June 2014 HUG - Continuuity Loom : Cluster Management

  • 270 views
Uploaded on

June 2014 HUG - Continuuity Loom : Cluster Management

June 2014 HUG - Continuuity Loom : Cluster Management

More in: Software , Technology
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Be the first to comment
    Be the first to like this
No Downloads

Views

Total Views
270
On Slideshare
0
From Embeds
0
Number of Embeds
0

Actions

Shares
Downloads
2
Comments
0
Likes
0

Embeds 0

No embeds

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
    No notes for slide

Transcript

  • 1. 100% Open Source, multi-tiered application infrastructure provisioning and management! Nitin Motgi Albert Shau @nmotgi @albertshau tryloom.io
  • 2. Our Problem 2 • Create variety of application infrastructure stack - Continuuity Reactor + Hadoop (CDH, HDP, Apache), Secured Hadoop, LAMP, Mesos, … • In Public and Private Cloud • Using our existing investment in SCM tools - Chef, Shell script, etc • Easy way to manage the stack - like add new services, manage configuration changes, scale up/down, etc.. • DevOps friendly, easy to manage and maintain • Resource constraints (1 Ninja support 20 Developers) - Ninja had no vacations or couldn’t get sick
  • 3. 3 Challenges Simply Hard Multi-tier applications are hard to provision, install, configure, monitor & scale and also have complex cluster life cycles. Lack of experience and repeatability Installation and Configuration require a lot of practice to tackle common pitfalls for configuring complex multi-tier application clusters, complexity prevents simple attempts at automation Increased organization bottlenecks Complexity limits number of people who can provision, install and configure complex multi-tier application clusters Decreased developer productivity Hard to support DevOps role within an organization without an easy to use self serving model No governance with reference architectures Difficult to model reference architectures in complex environments Multiple cloud environments Supporting multiple cloud environments is extremely hard.
  • 4. Our Assessment 4 • Numerous technologies available that solve parts of the problem • Open Source / Closed Source • We looked closely at Apache Ambari, Crowbar, etc… • Concluded that none at current state were solving our problem (Happy to talk about it after this talk)
  • 5. Introducing 5 A system for templatizing and materializing complex multi-tiered application reference architectures in public or private clouds. Designed bottom-up to support different facets of your organization - from developers, operations and system administrators to large service providers.
  • 6. Who can use it ? 6 CONTINUUITY LOOM TEMPLATES ON CLOUD DEVELOPER Create, provision and decommission clusters of complex, multi-tiered applications using a self-service model. SYS ADMIN Write Chef recipes, Puppet modules and shell scripts for application services. IT OPERATIONS Create and manage Continuuity Loom templates for multi-tiered application services. Configure compliance policies and enforce service constraints.
  • 7. Lifecycle 7 Create! Loom! Template Save! Template Publish! Template Expose! REST API Create! Instance of ! template on any cloud Reference ! Architecture! (Enterprise) Centralized! Administration! Control with! Multi-tenancy Supports Hybrid! Infrastructure! & Architecture
  • 8. Concepts 8 • Service! • Software component. ex: Namenode, Datanode, ResourceManager, etc. • Specify dependencies between services • Template! • Defines how a cluster should be created • Specify compatible services, configuration and constraints which determine service placement • Provider! • Creates and destroys machines. ex: AWS, Rackspace, Joyent, Openstack • Automator! • Performs service actions like start, stop, config, install, remove, init • ex: chef, puppet, scripts, etc.
  • 9. Architecture 9
  • 10. Cluster Creation: Solver 10 Template + args (size, etc.) Solver Layout services:{s1, s3} hw: large image: ubuntu12 x 1 services:{s2} hw: medium image: ubuntu12 x 4
  • 11. Cluster Creation: Planner 11 Layout Dependencies Planner Coordinator Task Queue ProvisionersProvisionersProvisionersProvisionersProvisionersProvisioners Plan
  • 12. Planner : DAG 12
  • 13. Planner : Linearize 13
  • 14. Planner: Expand 14
  • 15. DEMO 15
  • 16. 16 Thank you Nitin Motgi Albert Shau @nmotgi @albertshau http://github.com/continuuity/loom