11/08/2009 2:28 PM
GNS client vs. server mode
I'm not sure if it is right forum to ask question about GNS service - if not sorry for spam. My question is: What is the
difference between GNS in server and client mode. What I have already found gns in server mode is a database storing
all name bindings and gns client forwards all gns action to gns server. I have made some tests and found that node
running gns in client mode after process registration (name_attach()) creates directory \dev\name\global\process_name
(process_name – attached name). Could anybody describe me the core difference because gns client and server? And why
this directory is created which indicates that process was attached on this node?
P.S. Does anyone have some materials about QNX 6 message passing and gns mechanisms – (architecture, fault tolerance
mechanisms, etc.). I’m especially interested in consistency and replication mechanisms used during gns operations on
multiple gns servers.