The CentOS wiki page says: 'If you are installing FreeNX on a remote server, you will also need to install a Desktop environment on the machine in order to run the remote session.' Its never explicitly stated, but does the client also need to use the same DE? (I wince as I ask this because noMachine is also available for OSX. What's the deal?Ģ) I've selected the option to use the KDE desktop in the nxclient connection options. Why then does the CentOS wiki page tell you to add the private key copied from the server via ssh? I read somewhere, the private key is not shared and remains on the local machine. Now to a few simple and obvious questions:ġ) Here auth method says publickey. NX> 285 Enabling skip of SSH config files The NX service is not available or the NX access was disabled on host 192.168.2.8

Following the NoMachine setup for client and the CentOS FreeNX setup for client pages I've added the key, selected the desktop KDE (because its installed on CentOS-more on that below) and what I get is a kick in the. Nxclient launces on Arch box to give me access to the NX advanced setup. Local/nx-all 3.5.0-6 // NoMachine nxclient, nxnode, and nxserver in one package Local/libxcomp 3.5.0.21-1 // NX X compression library On Archlinux I've built the necessary package and installed: I regularly connect to this server using SSH via the default port, so there's no firewall issue with SSH. And at the end it suggested some fixes to ssh_config, hosts.allow, and iptables which I dutifully performed. This produced a bunch of ugly errors suggesting I wasn't going to be able to print and such. # nxsetup -install -clean -purge -setup-nomachine-key I used the nxsetup command I gathered from a 2009 forum archived post: On my CentOS6 server I've used yum to install NX and FreeNX: The CentOS FreeNX wiki makes it sound so easy.

Nomachine client 3.5