tfreeman at mcs.anl.gov
Wed Aug 9 11:52:26 CDT 2006
On Wed, 9 Aug 2006 16:03:15 +0100
"Dharanikota, Sai Srinivas" <sai.srinivas.dharanikota at intel.com> wrote:
> Could you please give pointers to the following questions?
> 1. Does the Workspace Service have the capability of managing more
> than one Workspace in TP1.1.1?
Yes, it has since the first release.
> 2. When does the Workspace service clean the virtual hard disks
> (VHDs) present on the physical machine? If you take the scenario where
> we keep on deploying Virtual Workspaces on the physical machine, at one
> point of time there will be no disk space on the physical machine to
> create a VHD.
In TP1.x, the service does not deal with image stage-in or stage-out. The TP2
We should discuss making the default action to delete the image in the case
where there is no stage-out (if there is stage-out, we obviously need to wait
until that is over). It didn't make sense in the TP1.x series because we were
leaving staging entirely up to the client to accomplish out of band.
In the future we will support an optimization where a VM can be constructed
from various signed partitions (for example a base image, application
partition, and vm/client-specific partitions). In that case, knowing when it
is OK to delete something needs more code because we a) want to make sure there
are zero users of the file and b) in many cases want to leave a cache of files
(the base images would be a good candidate) if space permits.
More information about the workspace-user