Direct VM & Baremetal discovery through Openstack api

Completed

Comments

5 comments

  • Avatar
    Arjen Annes

    Totally love that suggested addition, but plz also make it discover like on vmware and not as it does now on cloudstack. Discover the virtual machine names, instead of the internal ids. And have the virtual subtype set correctly from the start instead of mention it being an internalVM.

    0
    Comment actions Permalink
  • Avatar
    REDI TECH

    I was told we should be seeing this feature soon and its in QA.  Will this be released this year?

    0
    Comment actions Permalink
  • Avatar
    Raj Jalan

    Hi Ryan,

    OpenStack integration has been released in v8.3.2 in the 64 bit version of Device42. 

    http://blog.device42.com/2015/11/announcing-general-availability-of-the-64-bit-image/

    This was done a long time ago and we missed the announcement as were waiting for 64 bit release. Will add to the release notes now :)

    Thanks & Regards,

    Raj

    0
    Comment actions Permalink
  • Avatar
    Arjen Annes

    -    As it’s on per project basis be able to assign the customer to be set
    -    Provide an option to either register virtual machines under their id, or under their name (with the option to include the other in the alias). This to prevent naming conflicts as double names can be created if they are in different projects.
    -    Provide an option to run a LifeCycle Event instead of the single option to delete the unfound instances
    -    Provide an option to remove the dns-domain to get only the hostname registered
    -    Be able to choose the virtual subtype to set for the virtual machines (or set automagically not to InternalVM)
    -    Be able to set a tag per auto-discovery job
    -    Provide an option to import the networks (though might be tricky as customers are able to create their own networks and name them whatever they want)
    -    Maybe from the image name determine which operating system is installed (though that info could be gathered later on)

    Items working well:
    -    It gets the basic information
    -    It gets the vCPU, memory and disk sizes
    -    It picks up the ip, though that might become a challenge as within the various projects customers can possibly assign the same ip-addresses to their vms (though in closed environments then)

    0
    Comment actions Permalink
  • Avatar
    REDI TECH

    Feedback/Feature Requests:

    • Better Extended Debug output.  Why did the job fail?
    • Cloud autodiscovery jobs need a status column similar to network.  We should be able to see what is running without going into the job.
    • This may be possible with a quick report, but ability to see available/consumed compute for each project based on selected resources.
    • Bare Metal/Compute nodes are not being identified as physical devices. (I also have duplicate record issues, but looking into that further.)
    0
    Comment actions Permalink

Please sign in to leave a comment.