Forum / NoMachine for Linux / Connection reset by peer
- This topic has 9 replies, 2 voices, and was last updated 3 years, 1 month ago by fisherman.
-
AuthorPosts
-
September 24, 2021 at 16:23 #35433apr207Participant
Hi,
Sorry to bother you, but I keep having the NoMachine (version 7.4.1) connection reset by peer. I tried removing the ‘Use UDP flag’, similar to this suggestion:
https://forums.nomachine.com/topic/connection-reset-by-peer-2However, it still has the message ‘connection reset by peer’. I’ve attached the server log (log.tar.gz) and the client log (nx.zip) and the No Machine is running on Windows 10 (client) and Ubuntu Linux 16.06 (server).
I wondered if you could suggest anything?
Kind regards,
AndyAttachments:
September 24, 2021 at 18:00 #35442fishermanModeratorPlease update your server version to latest 7.6.2.
If you still reproduce the problem, please enable debug by executing
sudo /etc/NX/nxserver --debug --enable all
.
Then reproduce problem, and collect logssudo /etc/NX/nxserver --debug --collect
After collecting logs you could disable logs again by running
sudo /etc/NX/nxserver --debug --disable all
September 26, 2021 at 19:30 #35449apr207ParticipantHello,
Thanks for your reply. I have done exactly what you said, I am now using version 7.6.2.
I have also attached the logs.
My problem is interpreting the logs. Can you see what the problem could be?
Kind regards,
Andy
Attachments:
September 27, 2021 at 14:22 #35462fishermanModeratorHi,
Based on the logs you send us, nxserver process is still connected and we dont see problem reproduced.
Did you reproduced problem with connection reset by peer ?If yes, to make sure we have correct logs, please reproduce issue possibly few times again and send us player and server logs.
September 27, 2021 at 14:31 #35463apr207ParticipantHi,
Thanks, I just wanted to check, the NoMachine player is running on Windows 10, there is only the problem “connection reset by peer” viewed on the player side, the server it seems remains running. The server of NoMachine is running on Ubuntu Linux 16.04 LTS.
Could you tell me how to obtain the player logs on the Windows 10 NoMachine?
Kind regards,
Andy
September 27, 2021 at 19:12 #35473fishermanModeratorCollect server side logs automatically
More info is written under paragraph “Fourth Step: Collect Client Side Logs (if requested)”
On Windows the .nx folder is hidden. To find it, write the following in the address bar or the file browser:
%USERPROFILE%\.nxSeptember 29, 2021 at 17:33 #35515apr207ParticipantHi,
Thanks for your advice, I had disconnections at 14:34:11, 14:43:25 and 14:52:46 on Wed Sep 29 2021. It showed connection reset by peer on the player side three times. I have checked that I have the correct logs as per your instructions.
Are there any errors that correspond with these disconnections on the player (nx.zip) and server (NoMachine…zip) side?
Kind regards,
Andy
Attachments:
October 1, 2021 at 15:45 #35539fishermanModeratorFrom the logs we noticed that there is a problem that one process is hanging. With logs provided, the problem is hard to debug, to handle it we would require more data.
Please set in the /usr/NX/etc/node.cfg:
SessionLogClean 0
Also it seems, that master agent hangs in this case, so please add this line at the bottom of the /usr/NX/etc/node.cfg:
dumpCoreOnHang 1
or useecho "dumpCoreOnHang 1" >> /usr/NX/etc/node.cfg
This will cause the server to kill node with signal 11 instead of 9, so please make sure the system can handle core dumps. After this please restart the nomachine server:
/etc/NX/nxserver –restart
If the Agent hangs, the system produce core report for nxnode process. We would require the backtrace of all threads. The instruction is available here:
AR09L00810 – Using GDB to debug a core file on LinuxMake sure that debug, is enabled as per instruction Reproduce the problem and send the server and node logs as mentioned in the previous post as well as the debug traces from the core dump.
October 3, 2021 at 17:03 #35545apr207ParticipantThanks for your message. So this is what I did:
1) In /usr/NX/etc/node.cfg I set SessionLogClean 0 and dumpCoreOnHang 1
2) I restarted the server: /etc/NX/nxserver –restart
3) I am using Ubuntu 16.04 so I enabled core dumps like this: sudo service apport start
4) sudo /etc/NX/nxserver –debug –enable all
5) I reproduced the problem at least three times
Disconnected and re-Connected 14:54:31 on 3rd Oct 2021, 15:03:26, 15:12:25, 15:53:16
6) sudo /etc/NX/nxserver –debug –collect
7) sudo /etc/NX/nxserver –debug –disable all
8) I then stopped the core dumps: sudo service apport stop
9) Reverted SessionLogClean 1 and removed dumpCoreOnHang
Unfortunately no core files were created in /usr/NX/etc
October 4, 2021 at 16:46 #35568fishermanModeratorCrash is dumped by the OS, and depending on your configuration core file can be located in your home folder or /var/crash/ folder.
I noticed that uploaded logs exceeded upload size, so please can you email them at forum[at]nomachine[dot]com.
-
AuthorPosts
This topic was marked as solved, you can't post.