I have a volume with replication factor 2 on 2 nodes. So that each brick contains its own copy of all files... Is it possible to avoid using glusterfs client and write/read to/from volume directory directly? It seems to be working exactly as if I mounted it like this:
mount -t glusterfs IP:/mnt/my-volume somewhere... Whatever I write to the
my-volume directory on host
A I can see in the volume directory on host
Or can DFS be accessed exclusively via the glusterfs client and any other means probably lead to unexpected behavior?