If a volume is mounted many times from a single client, every mount should be considered a client to the volume. i.e if there are 4 `glusterfs' processes, they should still be considered 4 clients. Currently the number of clients is shown as 1. Gluster Clients : 1
Yes this is how it works. the clients list is added to a python set, so only unique client IP's remain. The idea was to provide admins with a view of how many client machines are connected to volume/cluster - basically so they can understand the environment i.e. 4 node rhs cluster supports 100 client machines. If I don't do this, one physical client will shows as n connections so the admin has no idea how many 'real' clients the rhs cluster is supporting. Make sense?
Can be closed as NOTABUG?
Cool. Can be closed. I'm closing it.