Forum Replies Created
-
AuthorPosts
-
Tor
ParticipantHi. Can you attach a screenshot of the notification?
Tor
ParticipantWhat’s slightly odd is there seems to be some interaction between the client and the server before this happens… the screen appears very briefly to have the preferred resolution… then it flickers and goes back to the higher resolution. I can upload a video somewhere if that helps to illustrate, but it’s bigger than the 1024k allowed by this forum.
Oh, I see, that means the client tries to set the resolution but Ubuntu is changing it back to the previous one. Are you connecting to the Ubuntu physical desktop or creating a new virtual desktop?
If you have the file.config/monitors.xml
in your home directory, can you try to rename it, logout, login and try to check if Ubuntu keeps the custom resolution when the client connects?Tor
ParticipantHi again. When the input stops working, you can disconnect and connect again the keyboard (for example turn it off/on or detach/attach the cable, if wired) to restore its functionality.
I confirm the fix is ready to be released, so we’ll publish it as soon as the other major changes are ready.Tor
ParticipantHey there, thank you for your reports!
For the first problem, can you try to enable the “use custom resolution” box and set the W 1280 and H 1024 resolution? This should work around the problem while we debug the reason for the ignored resize request.
For the second problem, we’re aware of it and we implemented a fix that will be included in the incoming release. I don’t remember if there was a method to work around it, I’ll talk with the team and let you know.
Tor
ParticipantOn the main window, when you right-click on a client machine, the selection marker rapidly moves up and down at a high frequency.
Can we exclude ghosts? 🙂 Can you please confirm if you’ve this popup issue on the client host, or in a NoMachine client running inside a NoMachine session? Also, in the former case, what is the Linux distribution version? Kernel version is important, but since we’ve to try to get a configuration as close as possible to the one you use, it is way better if we start from your same distro version.
March 6, 2025 at 10:41 in reply to: ED25519 and ECDSA for NX protocol produce “Authentication Failed” error #52124Tor
ParticipantAll’s well that ends well! 😀 Kidding aside though, there must be a reason somehow connected to your ssh tools and I would not exclude that it’ll happen again in future. We’re here to assist you in that case. Long live the keys!
March 5, 2025 at 11:20 in reply to: ED25519 and ECDSA for NX protocol produce “Authentication Failed” error #52094Tor
ParticipantFor some reasons, the key attached to the post is not valid, I’m getting these outputs when testing it with openssh 9.6p1
ssh-keygen -l -f testkeyfornomachine 256 SHA256:eWl6FVVBI16fflrvux+rY0JgcV0bQmeKaTdktJi0RGQ “TestKeyForNoMachine” (ED25519) ssh-keygen -y -f testkeyfornomachine Load key "testkeyfornomachine": invalid format
If you run the same commands on your end, do you get anything different?
March 4, 2025 at 19:43 in reply to: ED25519 and ECDSA for NX protocol produce “Authentication Failed” error #52085Tor
ParticipantWhile editing the Trouble Report with the recommendation submitted by @neatchee I noticed that there are some additional questions we can ask to collect details for the debugging, so here they are:
- Besides the key type security concerns, we assumed that the RSA key login is still working correctly. Could we get a confirmation?
- Where are the ED25519 and ECDSA key files located on disk? Did you maybe try to move them in different system folders (e.g. $HOME, tmp/, .nx/) to verify if anything changes?
- Are key access permissions correct, so both folder and file can be accessed by the user starting the player process?
- Can you share any of the key pairs not working, so we can test them in our labs? You can contact us at forum[at]nomachine[dot]com.
Thank you for helping us to debug this mysterious issue!
Tor
ParticipantInteresting, process could not even create the logs directory. Please run the command
/usr/NX/bin/nxplayer --log stderr
in a terminal, let me know what you see there as output.Tor
ParticipantUsually aborts are tracked also in client logs, so we’ve a good chance to find useful hints there. Could you pack your
$HOME/.nx
folder and send it to forum[at]nomachine[dot]com, please?Tor
ParticipantHi. There are no known issues like the one you’re experiencing, so we can troubleshoot it together by starting from a check of
journalctl
, it will help to verify if the client is segfaulting. Try to start it, then runjournalctl -r
and check if the processnxplayer.bin
was closed because of a segfault. Copy the line and paste it here in the thread.Tor
ParticipantHi. Are you using the touch or the pointer mode? Is the issue occurring in both cases? Can you try to open Display settings and enable the viewport mode, then verify if taps are correctly interacting with the desktop?
Tor
ParticipantHi biglia.
The version number 10.0.19041 was extracted from the dump you sent, and while we know that Microsoft has unclear rules about the DLL versioning (for example on all our full updated 22H2 machines, ntoskrnl has different versions!), we preferred to suggest to update your machine because this versions mismatch makes hard to compare systems and detect reasons for different behaviours, most of all when they occur in a deep system layer.
Troubleshooting your problem has been quite painful, as the NoMachine client process is closed before it can start doing anything meaningful: the Workstation Service DLL is loaded in a chain of dependencies and, when trying to do its stuff, raises an exception and causes the process to terminate. The exception description is “The network address is invalid”, and there is nothing the client did up there to cause an error like that.
After all our attempts to reproduce the issue and to progress in the debug through static analysis, we can certainly say that this problem is completely out of our control. A possible reason why it happens with NoMachine only (at least from what you’ve seen so far) is the sequence of DLLs loaded that badly interact with one or more configurations and third party services on your machine.
We tried to replicate your environment with WSL and WSL2, by publishing and accessing multiple resources on the network to extensively use Workstation Service, but we never succeeded to reproduce the same issue.
I remember you said that one of your machines with the same configuration works correctly. You could try to compare the services and configurations, and check event viewer to look for anything useful about wkscli. As of WSL, a good starting point could be the issue tracker on github. If you’ll find anything that you think valuable to be shared, we’ll be happy to know it. 🙂
Tor
ParticipantDo you need that the client machine runs also the server? If you do not, please try to shutdown and disable the server (Quit NoMachine service from the tray menu, choose to not start it again) and verify if the configuration is correctly moved.
Tor
ParticipantSome questions:
1. What do you mean with “server client” and “player client”?
2. Where are you changing the default path for logs, on client, on server, or both?
3. What is the product you’re using on client and on server?
4. How do you launch the client? -
AuthorPosts