When writing one large file to a 3 node Gluster configuration which has 6 bricks and a "Distrubuted" replica count of "2", only 2 out of 3 nodes will show any kind of hard drive activity.
Also only two nodes will show space utilization, the third node will show "0%" utilization.
One large file will be split among only two bricks as a mirror copy.
You would not be able to fill up the entire Gluster volume with a single file, that is not how Gluster with a "Distributed" replica count of 2 works.
To get the full performance and capacity out of the Gluster Volume with a "Distributed" replica configuration you will need to write/read multiple files and if possible from multiple hosts.
If you do have a configuration that needs to be a large single file spanning multiple nodes, you would want to look at providing that via our Ceph Scale-out Block style configuration or possibly via the Gluster Dispersed Erasure Coding mode which would allow for that sort of use case.
Additional detail on Gluster and Ceph configurations can be found below: