Cannot create virtual xdm session to Solaris 10

[termmachine]

Hi,

I set up a new Terminal Server 6.0.78 (also tried last 6.4 version) on Suse Linux Enterprise Server 12 SP4 and want to create a new XDM direct query session to an Solaris 10 Server with CDE, but it fails with message “Error: Cannot connect to NoMachine monitor.”

I also tried a XDM Session to an Ubuntu Server and this works. Moreover direct in a Suse terminal the command X -query 10.203.228.170 :1 also works, and i get the remote login screen instantly.

So I need your help please.

Attached you will find my complete configs and all logs I found.

Thank you!

[Britgirl]

The logs didn’t attach, most likely because they were too big. Please send them to forum_at_nomachine_com making sure to reference your topic. Thanks!

[termmachine]

I sent the attachments to forum(@)nomachine.com

Cant attach them here.

Thanks

[Mth]

Hello.

Unfortunately there are not really any clues in the logs, nothing specific, so we would require
additional help to solve this problem.

It does not seem the problem is connected directly with the target Solaris OS, but it is local.

First there is the failure of client –monitor, could you provide with logs from the nxclient?

For the session in your previous logs they will be in the

/home/domain/pozard/.nx/F-M-<hidden hostname>-13002-1B89DA468E176ABD25A2E2829A087C7F

then we would need agent logs, for the same session they should be in:

/usr/NX/var/log/node/F-C-<hidden hostname>-1002-FE980136BA5DA062CC519F72B5CBC494

Then some questions:

1. It seems that there is Kerberos configured on this system, could you check if there are any errors for the user ‘pozard’?

In the logs we have an error regarding that ticket, but the user directory access seems to be working so it may or may not be related.

2. When you say the XDM to Ubuntu server works, did you also use user ‘pozard’? If no, please check what happens.

3. When you try the failed session, are there any other NoMachine sessions for user ‘pozard’ running?

/Mth

[termmachine]

Hi,

thanks for your reply and I sent the additional log files to forum(@)nomachine.com again.

To your questions:

1. It seems that there is Kerberos configured on this system, could you check if there are any errors for the user ‘pozard’?

I use Kerberos in combination with samba and winbind, because my users should be able to use the credentials from our companies active directory (windows domain). The authentification works, but I will check the logs.

In the logs we have an error regarding that ticket, but the user directory access seems to be working so it may or may not be related.

User directory access works, I have tested that.

2. When you say the XDM to Ubuntu server works, did you also use user ‘pozard’? If no, please check what happens.

Yes, I also used the pozard user. And tried different users as well. same problem.

3. When you try the failed session, are there any other NoMachine sessions for user ‘pozard’ running?

I tried it with no sessions besides the solaris xdm session and i tried it with other sessions (to other systems) openend besides. the xdm to solaris always fail with cant connect to NoMachine monitor.

 

Thanks!

[Mth]

Hello.

It seems that we need to do more debugging.

Please edit the file:

/usr/NX/bin/nxclient

and add following lines:

NX_MONITOR_DEBUG=1
export NX_MONITOR_DEBUG

before the last line, so the end of this file look like:

NX_SYSTEM=”$NXPath”
export NX_SYSTEM

NX_MONITOR_DEBUG=1
export NX_MONITOR_DEBUG

exec “$NXPath/bin/nxclient.bin” “$@”

Then please reproduce the problem with connection to Solaris and send us the nxclient
logs. This logs will be under the:

/home/domain/pozard/.nx/

directory, where will be a new F-M-* directory created.

/Mth

[termmachine]

Hi,

thanks for your reply, I sent the new logs via E-Mail after enabling of debugging.

Thanks!

[Mth]

Hello.

It seems that NX have a problem with connecting to the Solaris Xserver using query
option.

Could you try to connect using “Get a list of available X desktop managers” and check if
you receive the list window and if you can chose the working desktop manually.

/Mth

[termmachine]

Hi,

I am able to get a list of available X desktops and can choose one. But when I want to connect to one, I get a black screen and after a defined time I get kicked out from webplayer. This happens with all of the available desktop managers from the list.

This is the nxclient logfile from one try:

28078 28078 08:40:01 777.753 Main: Enter main function.
28078 28078 08:40:01 777.885 Main: Initialize file ownership.
28078 28078 08:40:01 778.104 Main: Completed file ownership initialization.
28078 28078 08:40:01 778.138 Main: Update parameters.
28078 28078 08:40:01 778.151 Main: Completed parameters update.
28078 28078 08:40:01 778.190 Main: Creating the application in GUI mode.
28078 28078 08:40:01 778.203 Main: Set client environment.
28078 28078 08:40:01 778.229 Utility: Going to set process environment.
28078 28078 08:40:01 778.239 Utility: Environment personal NX directory is [/home/DOMAIN/pozard/.nx].
28078 28078 08:40:01 778.246 Utility: Environment system NX directory is [/usr/NX].
28078 28078 08:40:01 778.251 Utility: Not using parameters.
28078 28078 08:40:01 778.258 Utility: Checking the file [/home/DOMAIN/pozard/.nx/config/player.cfg].
28078 28078 08:40:01 778.290 Utility: Using [/home/DOMAIN/pozard/.nx/config/player.cfg] as config file.
28078 28078 08:40:01 778.297 Utility: Using System NX directory.
28078 28078 08:40:01 778.308 Utility: Reading Personal NX directory.
28078 28078 08:40:01 778.314 Utility: Using Personal NX directory.
28078 28078 08:40:01 778.339 Utility: Loading the global settings.
28078 28078 08:40:01 779.287 Main: Starting slave server.
28078 28078 08:40:01 779.306 MonitorService: Start slave server.
28078 28078 08:40:01 779.419 MonitorService: Create the slave server.
28078 28078 08:40:01 779.431 MonitorService: Creating the slave server instance 11315584.
28078 28078 08:40:01 779.689 MonitorService: Starting the slave server instance 11315584.
Info: Slave server running with pid 28085.
28078 28078 08:40:01 779.709 Main: Initialize qt application.
Info: Listening to slave connections on port 25002.
08:40:01 780 QApplication: Initialize qapplication.
08:40:01 780 QApplication: Create QX11 data.
08:40:01 780 QApplication: Initialize QX11 font config.
08:40:01 786 QApplication: Completed QX11 font config initialization.
08:40:01 786 QApplication: Initialize XRender.
08:40:01 786 QApplication: Completed XRender initialization.
08:40:01 786 QApplication: Initialize application class.
08:40:01 786 QApplication: Completed application class initialization.
08:40:01 786 QApplication: Install default error handlers.
08:40:01 786 QApplication: Completed default error handlers installation.
08:40:01 786 QApplication: Check command line parameters.
08:40:01 786 QApplication: Completed command line parameters check.
08:40:01 786 QApplication: Connect to X server.

Is there a chance to get this fixed in a new nx version update? The remote connection via webplayer was the main reason to buy this product, but if this does not work, the product has been purchased without benefit for our company.

Thanks!

[Britgirl]

Hi, we’re reproduced the problem and are investigating further to find a suitable fix.

At this point, it might be an idea to open a support ticket in your customer area (you mention that you’ve purchased a subscription), that way you’ll also be able to submit more information without worrying about it being made public as well as get access to any hot fix we make.

 

[Mth]

Hello

Please check the following workaround for this problem:

Edit the:

/usr/dt/config/Xconfig

configuration file on the Solaris system so the Dtlogin*grabServer field
is set to False and is uncommented:

Dtlogin*grabServer: False

then restart the cde-login service for example by running:

# svcadm restart cde-login

Then try XDM session again.

/Mth

[termmachine]

Hello,

thanks for that workaround. With this change XDM on Solaris is possible and I can login.

Is the workaround the final solution, or will this be fixed in an future update of Terminal-Server?

Moreover I found another bug, one of my programs on Solaris needs a Font-Server and if I set our Font-Server in node.cfg (#AgentFontServer) this isn’t working.

When opening the program on Solaris, the NoMachine Client crashes with I/O Error (5)

Thanks!

[graywolf]

We’ll fix it in NoMachine.

About the font server issue, check the current font path in the remote session by command xset q, then try add font server to path, e.g.:

xset fp+ unix/:7100 (font server on localhost, Unix socket)

xset fp+ tcp/10.0.0.1:7100 (font server on remote host 10.0.0.1, TCP)

[termmachine]

Hello, thanks for your answer.

I solved the font problem with automounting (via NFS) the specific fonts directly on my Terminal Server.

 

[Author]

Posts