Spraying/copying between different size cluster appears to be mangling the size in the target file

The example seen was a spray from a 50-way foreign source file, to a 400-way target in the local environment.

The resulting 400-way file appeared to be valid and could be used, but running File.Std.VerifyFile() on it detected a mismatch in the meta per part size info vs the actual size on disk.

On closer inspection (using daliadmin <dali-ip> dfsfile), looking at the meta info, it appears that the target file had copies of the source per part @size info, which is wrong given the target par part info should be different (smaller), given it was split approximately 8 ways per part.