I have a raspberry pi running arch linux connected to the TV and want to run commands on that screen by SSH'ing from my Ubuntu machine elsewhere.
I have tried running the command SSH -Y root# and starting xclock as a test. However I get the following error:
No protocol specified
Error: Can't open display: :0
I then tried running "export DISPLAY=:0" on the arch machine, but this doesn't solve the issue.
I have X running fine and can see it on the TV and I have edited the file /etc/ssh/ssh_config to allow X11 forwarding.
Anyone know what I am doing wrong?
Many thanks
The ArchLinux wiki has some good instructions on setting things up. I struggled for a while until I found them, but for me the key was to set up SSH forwarding within sshd_config (most distros seem to enable this by default, but Arch doesn't).
Once you've enabled X forwarding in SSHD, I also found it helpful to install the dummy video driver for Xorg ("pacman -S xf86-video-dummy"), which gets rid of irritating errors about missing the RANDR extension. This might not be necessary if you already have it running a real X server locally on the Pi - personally, I'm running mine completely headless.
I've found that I don't have to worry about setting up the DISPLAY variable, or xhost, when running with 'ssh -X' from an iMac. I'd expect Ubuntu to behave the same.
looks like an authentication problem. The answer used to be
xhost +
on the console of the X display. Might help.
You probably want to set
export DISPLAY=[IP_ADDRESS_OF_ARCH_BOX]:0
on your Ubuntu box, then when you run xclock from your Ubuntu box, it'll forward the screen to the Arch box.
Related
I'm connected to a remote Linux server via ssh, and I'm trying to execute ImageMagick's display command. I know that the problem is probably not on the server side, because when I connect from different machines I don't get the same problem. And I installed XQuartz on my local machine as well as connected using the ssh tags -Y and -X (neither fix the problem) and I'm still getting the following error. My computer is a MacBook Pro running El Capitan. Any idea?
display: unable to open X server `' # error/display.c/DisplayImageCommand/426.
Thanks!
I think the trick is to start XQuartz in the background on your Mac before you do anything and allow remote machines to connect.
So, on your Mac, in Terminal, start XQuartz:
X &
Now right-click the XQuartz icon in your Dock and click:
Applications -> Terminal
and in the Terminal that appears, type:
xhost +
Later, when you have it working, tighten up the security on the xhost command by being a bit more specific with IP addresses.
I'm running a VM in Azure with Ubuntu 14.04. I've installed XFCE and use X2GO to remote desktop to the machine.
I've downloaded the VS Code and unzipped the files. When I run Code from a terminal I get the following error:
WouterDeKort#UbuntuDev:~/tools/web/visual-studio-code$ ./Code
bash: cannot set terminal process group (-1): Inappropriate ioctl for device
bash: no job control in this shell
WouterDeKort#UbuntuDev:~/tools/web/visual-studio-code$ [ , [Error: channel closed] ]
Any idea what I'm doing wrong?
DISCLAIMER ;-) I'm a total Linux newby. I've just installed Linux for the first time today and I have no idea what I'm doing ;-)
The work-a-round is in issue 3451. After running this, I was able to launch it from Debian Jessie with LXDE over X2Go.
sudo sed -i 's/BIG-REQUESTS/_IG-REQUESTS/' /usr/lib/x86_64-linux-gnu/libxcb.so.1
I have the same problem using X2GO and LXDE. Works okay (but slowly) using MobaXterm so apparently this is a problem specific to X2GO.
For anyone who still wonder what is the best choice if you want to start Visual Studio Code on a headless Ubuntu - the winning combination is:
Ubuntu Server + xfce + nomachine remote desktop
VS Code starts immediately without any hacks, you can set your resolution however you like, and nomachine is blazingly fast.
At work I get a choice of two laptops to use, Windows or Mac. Being a linux guy, I thought the MacBook Pro will be a good fit.
So I bring the MacBook Pro laptop home, and realize it doesn't have a right-ctrl key. And the bottom panel is annoying as hell. So I say, no problem, I'll sort all that out later, for now I'll use it as a server and remotely ssh into it from my crispy Xfce workstation environment. Well...
I've spent about a day trying to figure out:
How to ssh into the mac from my Fedora workstation, run emacs and have it show up (XForwarded) on my linux workstation ? You know like this:
I've installed XQuartz and changed /etc/sshd_config and /private/etc/sshd_config with:
X11Forwarding yes
XAuthLocation /opt/X11/bin/xauth
Mind you, xeyes works, and so does xclock, yay !!!
Apparently X forwarding is working okay, since other gui apps are showing up remotely. But as for Emacs, foggetaboutit.
The default Emacs from /usr/bin/emacs just runs it in -nw mode. Then I've installed the latest Emacs 24.3 (into /Applications/Emacs.app/Contents/MacOS/Emacs). This time, if I'm physically logged into the laptop (i.e. from the laptop keyboard) running Emacs from ssh shows up on the laptop's screen !!! WTF? If I logout of the laptop, then I get:
_RegisterApplication(), FAILED TO establish the default connection to the WindowServer, _CGSDefaultConnection() is NULL.
I'm at my wits ends. Any ideas why other X apps work? Anything special about Emacs?
The Emacs.app binary is built to use Mac OS X's window system (we call it "ns" around here), which has nothing to do with X11, so you can't forward it over the network to another host. Instead, you want to install another Emacs on your Mac, which is built to use the X11 window system (so you can use it locally via Xquartz (which lets X11 clients use the native Mac OS X display), or you can use it remotely like any other X11 client).
The easiest way is probably to built it yourself: get the emacs-24.3.tar.gz source code and compile it with ./configure --with-x; make.
With Mac Ports x11 or gtk variant has to be selected, like this:
sudo port install emacs +x11
or that:
sudo port install emacs +gtk
hope everyone is doing well.
Got an issue I would like to pick the great brains in this forum about.
So, I have many Debian machines I often remote into using Teamviewer and a Windows machine, which works out great; however Teamviewer crashes for what appears to be no apparent reason. When such happens, normally I will PuTTY into the suspect Debian machine from a Windows 7 machine and issue the shutdown -r now command. When the debian computer reboots, so does Teamviewer and the suspect Debian machine is now visible in the "My computers" list on my Windows machine.
My question is this, why cant I just type teamviewer from the command line while PuTTYing into the Debian machine and have Teamviewer reopen on the Debian machine, making it visible in "My computers?" When I execute teamviewer command from PuTTY command line, I see the verbose output of teamviewer opening, but the machine never becomes visible in "My computers."
Now, I know it is possible to start Teamviewer remotely and have it show up back in 'My computers" list. I can do it via NoMachine; I simply connect to the suspect Debian machine and double click on the Teamviewer desktop launcher. However, NoMachine is very heavy and quite frankly, i just don't like it. I also know I can execute teamviewer via command line while sitting in front of the suspect Debian machine. PuTTY would be so much faster.
I have also tried VNC as well as X sessions. In both cases, i can open Teamviewer and it shows up in "My computers," but when I close the VNC of X, Teamviewer also closes and removes itself from "My computers."
Can someone help me figure out a way to restart Teamviewer via PuTTY please? Is there a way to open a desktop launcher via PuTTYs command line? Or is there a way to tell Teamviewer to execute in the suspect Debian machines active desktop session via PuTTYs command line?
Thanks all
I am using Mobaxterm(free version) on a windows 7 desktop to connect to a SUSE 11 Enterprise server on AWS. I am trying to display the xclock program on my xtrem client but I get an error saying 'Error: Can't open display:'. I have used the following syntax to set the display on the server:
export DISPLAY=<IP_addr>:0.0
SUSE 11 does not come with xclock by default so I had to download it and install it.
The hosts file on my PC has the localhost entry commented out, I am not sure if that would make a difference. Any ideas on how to debug this? Thanks.
Fixed!
Earlier I was just looking at xclock program's error msg. But when I scanned Mobaxterms client terminal's output, I found the following msg:
X11 forwarding request failed on channel 0
After some google hunting, found that one of the reasons this happens is when xauth package is not installed on the remote server. So, I checked and found that to be the case. This is the command I ran:
zypper in -name xorg*
This command tells you if the package is installed and if any dependencies exist. The package comes bundled with the xclock program. So zypper uninstalled the other xclock I had installed from another source and replaced it with the right version.
Link to package info:
https://www.suse.com/LinuxPackages/packageRouter.jsp?product=server&version=11&service_pack=&architecture=i386&package_name=xorg-x11-xauth
I also modified a file called /etc/ssh/sshd_config as root. The following lines need to be uncommented:
X11Forwarding yes
X11DisplayOffset 10
X11UseLocalhost yes
And X11UseLocalhost was changed to 'no'. I also changed my security group on AWS to let inbound traffic on port 6000. I am not sure if that matters.
After this mobaxterm automatically set my display parameter to localhost and I was able to run xclock on the remote server and see it on my local PC desktop.