Results 1 to 4 of 4

Thread: A few questions about virtualisation and reliability

  1. #1
    Join Date
    Oct 2013
    Beans
    1

    A few questions about virtualisation and reliability

    Hello,

    I am new to the virtualisation techniques and have a few questions.

    If I have several physical servers, is there a way of creating some kind of redundant VMs, so that if one server fails, the other VM will instantly do all the work of the failed VM?

  2. #2
    Join Date
    Mar 2010
    Location
    Metro-ATL
    Beans
    Hidden!
    Distro
    Lubuntu 12.04 Precise Pangolin

    Re: A few questions about virtualisation and reliability

    You can create clusters with VMs just like you do with regular systems.

  3. #3
    Join Date
    Jun 2006
    Location
    Houston, TX
    Beans
    1,120
    Distro
    Ubuntu 6.06

    Re: A few questions about virtualisation and reliability

    There are as many ways to do this are there are systems. In VMware, they call it vmotion, for example. Outside of visualization, you can use something like CARP. (You could do it virtually as well) Some poeple use puppit to monitor and restart a stuck VM, or start a different one.

  4. #4
    Join Date
    Mar 2010
    Location
    Metro-ATL
    Beans
    Hidden!
    Distro
    Lubuntu 12.04 Precise Pangolin

    Re: A few questions about virtualisation and reliability

    Quote Originally Posted by houstonbofh View Post
    There are as many ways to do this are there are systems. In VMware, they call it vmotion, for example. Outside of visualization, you can use something like CARP. (You could do it virtually as well) Some poeple use puppit to monitor and restart a stuck VM, or start a different one.
    Moving VMs between physical hardware has been supported by KVM and Xen for some time too. To my knowledge, all of them require a SAN/NAS setup for the VMs for this to work ... except extremely recent versions of ESXi which can use local storage on both and migrate VMs (according to their announcement at VM-World a few months ago).

    Moving a VM is not the same as clustering, however. If you have a service that is important enough to need clustering then nothing else will probably do. If a crashed service being restarted in a few minutes is acceptable, then the other alternatives are much easier to administer.

    For webservicey things, I like using the sticky DNS methods and having N+1 front-ends, then going to back-end systems with similar setups as much as possible. It really depends on the specific architecture of the solution.

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •