VMware: Same image should not be transferred more than one time on concurrent spawn operations

[Arnaud Legendre]

On concurrent spawn operations, the VMware driver transfers the same image into different temporary directories before getting into the cache: the same bits might be downloaded 20 times if 20 spawn operations happen at the same time for a given image (before being cached).
The goal here is to use a lock to only transfer one time the bits of the image.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/75551

[Joe Gordon (jogo)]

Is this bug still valid? The patch was abandoned.

[zhu zhu (zhuzhubj)]

This bug should be a valid bug. For the same flat image deployment in concurrent, it will transfer same image multiple times. Though deployment finally could get success, it will waste a lot of datastore disk spaces.

[Sean Dague (sdague)]

feature for performance enhancement