Glance

  1. Bug #1498645
  2. Comment #4

Comment 4 for bug 1498645

Revision history for this message
Sabari Murugesan (smurugesan) wrote :

Thanks Giri for filing the bug.

Yup, as I mentioned, the vCenters http read timeout needs to be increased because there may be brief periods of time without any data transfer as the ESX server is busing compressing the disk for streaming. The http pipeline looks something like this and unfortunately the whole operation is serialized :-

ESX Server --> nova-compute --> glance-api --> vCenter Server

Unfortunately, we can't fix this bug on glance's VMware store.