Josh Thompson updated VCL-66:

    Fix Version/s:     (was: 2.3)
> Reservations for newly created images may be assigned to management nodes or 
> computers on which the image can't run
> -------------------------------------------------------------------------------------------------------------------
>                 Key: VCL-66
>                 URL: https://issues.apache.org/jira/browse/VCL-66
>             Project: VCL
>          Issue Type: Improvement
>          Components: database, web gui (frontend)
>    Affects Versions: 2.0
>            Reporter: Andy Kurth
>            Priority: Minor
>             Fix For: 2.4
> After a user creates a new image, that user can create a reservation for the 
> image without having to configure any groups or mappings.  This is a good 
> thing.  The image is automatically configured to be a member of the 
> 'newimages-<username>-<user id>' group.  The 'newimages*' image groups are 
> mapped to the 'newimages' computer group.  As a result, reservations for new 
> images will be assigned computers in the newimages computer group.
> Although this usually works fine, the automatic mapping to the fixed computer 
> group could result in an image failing to load because the computers in the 
> group may contain different hardware than the image was created on.  
> A different management node may also be assigned than the one the image was 
> created on.  This could also result in an image failing to load if the 
> management node is running different code which doesn't support the image or 
> configured in a way which doesn't support the new image.  This may also cause 
> wasted image repository space if the user quickly discards the image because 
> the image would be copied to additional management nodes.
> This issue was created after encountering a limitation while I was creating 
> the initial Vista base image.  I needed to allow other users to create 
> derivative image from the base image but wanted all derivative images to run 
> from the same management node.  At the time, only 1 management node fully 
> supported the Vista code and code updates were being made frequently as the 
> bugs were worked out.  Because of this issue, other users' Vista images would 
> attempt to load on other management nodes and fail.
> It may be beneficial to at least limit new images to the same management node 
> that they were created on until specific image grouping is configured for the 
> image.  It would be much more difficult to have new images only run on the 
> same hardware.
> One possibility may be to have a management node to module mapping.  As more 
> and more components are modularized, we could achieve greater flexibility if 
> management nodes could be configured to support certain modules.  In this 
> example, I would have simply had to map the Vista module to the management 
> node which supported the code and added some computers to the newimages 
> computer group.

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: 
For more information on JIRA, see: http://www.atlassian.com/software/jira


Reply via email to