Hans, I wonder if this is the same issue that is described in bug report #1520 ( https://bugs.x2go.org/cgi-bin/bugreport.cgi?bug=1520). Can you try the workaround from that bug report?
Thanks, Adam
On Wed, Aug 25, 2021 at 6:03 AM Hans Peter Verne <h.p.verne@geo.uio.no> wrote:
Package: x2goclient Version: 4.1.2.2
Hello, devs!
I'm afraid this is not a very precise bug report. Any assistance in how to investigate this is appreciated.
We're running x2goclient 4.1.2.2 on Windows-10 machines, this has served us well up to now. With our new RedHat Enterprise Server 8.4 in production, we see the client crashing very often.
When the client crashes, it does so shortly after trying to log in to a new or existing session. The clients "show details" window pane is activated, but the client dies before anything can be seen there.
On the server, just these entries in the system logs appears:
Aug 05 14:42:56 mimi.uio.no sshd[2421324]: Connection from 2001:700:100:4028:9462:e3c7:21c5:ec1c port 49763 on 2001:700:100:118::101 port 22 Aug 05 14:42:59 mimi.uio.no sshd[2421324]: Connection reset by 2001:700:100:4028:9462:e3c7:21c5:ec1c port 49763 [preauth]
If I disable IPv6:
Aug 05 14:45:30 mimi.uio.no sshd[2421601]: Connection from 193.157.161.44 port 57619 on 129.240.118.101 port 22 Aug 05 14:45:33 mimi.uio.no sshd[2421601]: Connection reset by 193.157.161.44 port 57619 [preauth]
Note the absence of the "Fail password for ..." log entry. It looks like it never gets to even try authenticating.
There is no problem logging in with putty from the same client machine, both with IPv6 and IPv4. Neither have I had this problem (at least not so frequent) with the Linux client.
The server is running sshd and x2goserver from the RHEL8/EPEL repos, packages openssh-server-8.0p1-6 and x2goserver-4.1.0.3-9, respectively.
I eventually tried to run sshd -ddd on the server and catch the output. The typescripts are attached (slightly edited): one where the client crashed; and one where it didn't (but couldn't authenticate, bad password). I'm no expert on ssh, but it seems like the difference appears after the key exchange ("KEX done"), when the instance that crashed never reaches the bit with userauth-request.
I also tried "setenforce 0" on the server, i.e. disable SELinux, but the client still crashed.
As it is, I can't clearly provide a procedure to reproduce the problem, as it doesn't always happen.
If more information is required, or there are suggested steps to take, please let me know.
Thanks in advance, and best regards,
Hans Peter Verne -- IT-drift Geofag.
In 1934, Van der Lubbe was beheaded in a German prison yard. In 1967, a court in West Berlin overturned the 1933 verdict, and posthumously changed Van der Lubbe's sentence to eight years in prison. -- "Reichstag fire" on Wikipedia.
x2go-dev mailing list x2go-dev@lists.x2go.org https://lists.x2go.org/listinfo/x2go-dev
-- Adam Dorsey NOAA RDHPCS Systems Administrator Site Lead CSRA / RedLine Performance Solutions, LLC
NOAA NESCC 1000 Galliher Drive, Suite 333, Fairmont, WV 26554 office: (304) 367-2882 cell: (304) 685-9345 adam.dorsey@noaa.gov