using vsphere replication - "no space left on device" ... but there is plenty of available space?

[TedH256]

Replicating VMs as part of an SRM 5.1 implementation, and using vSphere replication to do it. 1GB link between sites. All recovery side data stores are NFS on NetApp with VAII plugin running, datastores on protected site are HP EVA fc.

The error on the protected site:

there is a storage issue with datastore path --  there is no space left on device or datastore path

But in fact there is plenty of space left (or so it appears).

How much free space must be available on a target lun in order to replicate successfully, and does anyone know of a good solid technical article detailing space requirements and the hbr technology that vSphere replication uses?

[TedH256]

I'm thinking that this question belongs in the vsphere replication forum, but there is so little activity there - anyone have a good detailed technical white paper on how replication works, and how it uses target datastore space, and what the hbr disk process is, and what effect having thick provisioned source disks has on the target datastore? (do the HBR disks take up a different amount of space depending on whether either the source or target are configured as thick or thin)?

In some cases it is looking like the HBR disks (there is one for each replicated disk) are equal in size to the source disks. In other cases it looks like they are a small tiny subset of the larger source size.

It would be really good to have a detailed description of all of this and how it is supposed to work