Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Cloud Computing in practice with OpenNebula ~ Develer workshop 2012

Cloud Computing in practice with OpenNebula ~ Develer workshop 2012

A short introduction to OpenNebula for newcomers, with a peek on distributed filesystem and monitoring tools.

Giovanni Toraldo

November 26, 2012
Tweet

More Decks by Giovanni Toraldo

Other Decks in Technology

Transcript

  1. Cloud Computing,
    in practice
    with
    Giovanni Toraldo - Sysadmin

    View Slide

  2. ● Keywords: Linux, Debian, Ubuntu, PHP, Drupal,
    Symfony, Open Source, pizza, beer, geek.
    ● OpenNebula contributor, currently working on
    Cloudesire at Liberologico.
    about.me/giovannitoraldo

    View Slide

  3. About OpenNebula
    OpenNebula.org is an open-source
    project developing the industry
    standard solution for building and
    managing virtualized enterprise data
    centers and IaaS clouds

    View Slide

  4. A bit of history

    View Slide

  5. Core features
    ● Virtual infrastructure management and
    orchestration (KVM, XEN, ESXi)
    ● External cloud connector (EC2)
    ● Interfaces for Cloud Consumers (EC2, OCCI)
    ● Capacity and performance management
    ● Monitoring
    ● Accounting
    ● User quota, ACLs

    View Slide

  6. Why not something else?
    ● Focusing on data center virtualization
    ● High perf core, scriptable drivers
    ● All-in-one package, batteries included
    ● Opensource and user-driven development
    ● API-agnostic

    View Slide

  7. Architecture

    View Slide

  8. Shared storage

    View Slide

  9. Shared storage: NFS or?
    GlusterFS: completely decentralized, no
    SPOF, FUSE client, developed by Red Hat:
    http://gluster.org - use ON shared driver.
    MooseFS: no native HA, proven stable,
    automatic balancing of resources:
    http://moosefs.org - use ON shared driver.
    Ceph: should solve a lot of problems, but
    actually only the block device is "stable":
    http://ceph.com - use experimental driver

    View Slide

  10. Monitoring
    ● Ganglia: near-zero configuration,
    distributed monitoring and graphing for
    HPC, clusters, and highly dynamic
    environments.
    ● im_ganglia driver can be used to off-load
    frontend from hosts resources polling via
    ssh.

    View Slide

  11. View Slide

  12. Monitoring
    ● Check_mk: use a local agent to gather
    metrics on every node, generate
    nagios/icinga configs to retrieve them and
    generate alerts when shit happens.
    ● HOST_HOOK and VM_HOOK can be used to
    dynamically attach/detach hosts and VM
    instances within Check_mk.

    View Slide

  13. View Slide

  14. It's time to party...

    View Slide

  15. View Slide

  16. Overview of the CLI Interface
    # Login as root/opennebula and su on oneadmin
    $ su - oneadmin
    # Currently configured hosts and their status
    $ onehost list
    # Currently available vnetworks
    $ onevnet list
    # Show leases and specific configuration of the vnetwork
    $ onevnet show 0
    # List available disk images
    $ oneimage list
    # List registered VM templates
    $ onetemplate list
    # Show template details
    $ onetemplate show 0

    View Slide

  17. Resources
    http://opennebula.org/documentation:rel3.8
    http://blog.opennebula.org/
    http://bit.ly/opennebula-3-book-packtpub
    http://marketplace.c12g.com/appliance
    http://dev.opennebula.org/
    http://opennebula.org/software:ecosystem

    View Slide