Failing to reconnect to an NX session with the error: "error in locking authority file"

I am able to start an NX session without any issues. But, when I 'disconnect' (leaving a persistent connection, as opposed to 'terminate'), I am unable to reconnect. Here is the error from /var/log/messages:

    Aug  5 19:05:41 praveen NXNODE-3.5.0-3[20703]: ERROR: NX> 596 ERROR: NXNODE Ver. 3.5.0-3  (Error id e77ECD3) [e77ECD3] Logger::log nxnode 2963
    Aug  5 19:05:41 praveen NXNODE-3.5.0-3[20703]: ERROR: NX> 596 ERROR: create session: run commands [e77ECD3] Logger::log nxnode 2963
    Aug  5 19:05:41 praveen NXNODE-3.5.0-3[20703]: ERROR: NX> 596 ERROR: execution of last command failed [e77ECD3] Logger::log nxnode 2963
    Aug  5 19:05:41 praveen NXNODE-3.5.0-3[20703]: ERROR: NX> 596 last command: /usr/bin/xauth -v source /home/pmohandas/.nx/C-praveen-1003-707A5CF0D1B5184E1EBEDC07BB4981EF/scripts/authority [e77ECD3] Logger::log nxnode 2963
    Aug  5 19:05:41 praveen NXNODE-3.5.0-3[20703]: ERROR: NX> 596 exit value: 1 [e77ECD3] Logger::log nxnode 2963
    Aug  5 19:05:41 praveen NXNODE-3.5.0-3[20703]: ERROR: NX> 596 stdout:  [e77ECD3] Logger::log nxnode 2963
    Aug  5 19:05:41 praveen NXNODE-3.5.0-3[20703]: ERROR: NX> 596 stderr: /usr/bin/xauth:  error in locking authority file /home/pmohandas/.Xauthority [e77ECD3] Logger::log nxnode 2963

I am on a RHEL6.1 x86_64 machine.

I tried running xauth -b quit prior to reconnecting to the NX session. No luck with that either.

Does anyone know how I cou开发者_Go百科ld get rid of this error? I find NX to be much snappier than VNC and would love to get this to work :) Thanks in advance!

I don't know the cause of this annoying error yet either, but a workaround is to ssh to the nx server machine and remove your ~/.Xauthority-c and ~/.Xauthority-l files. You'll lose any existing sessions, but you will still be able to connect.

Solution here from : http://forums.fedoraforum.org/showthread.php?t=223609

First I must say that I'm running my installation in a virtual machine but I think that doesn't matter.

Problem: I found out that the hostname I enter at the installation wasn't added to the /etc/hosts file. There were only the default entries for localhost/localdomain. So the DNS server (see /etc/resolve.conf which is generated by the network manager) couldn't resolve my hostname which is needed for xauth. You can check this by traceroute your hostname (note nslookup or host won't work because they ignore /etc/hosts).

Solution: I just added my hostname to the /etc/hosts file, checked it with traceroute and connected via nxclient from my windows host.

So adding server.craneworks to /etc/hosts should solve the problem for you.

The problem seems to happen when you have duplicate entries in your /etc/hosts. Here is the xauth bug: http://www.kerneltrap.com/mailarchive/openbsd-bugs/2009/3/8/5118304

Fixed the problem by removing the duplicate entries in /etc/hosts.