For the issue reported here I would propose starting with a simple solution that cleans up any resize disks during the virt driver destroy() call. The xenapi driver already does a similiar thing for rescue disks.
For the issue George is discussion I would look at adding a small store of instance->task mappings to each driver that it can use to clean up tasks during the destroy() process.
For the issue reported here I would propose starting with a simple solution that cleans up any resize disks during the virt driver destroy() call. The xenapi driver already does a similiar thing for rescue disks.
For the issue George is discussion I would look at adding a small store of instance->task mappings to each driver that it can use to clean up tasks during the destroy() process.