[slurm-users] error: _x11_socket_read: slurm_open_msg_conn: Connection refused

[Weijun Gao]

Hi all,

I compiled Slurm 18.08.7, 19.05.5 and 20.02.0 on Ubuntu 18.04 server.

X-Forwarding works fine with 18.08.7.

* (19.05.5 / 20.02.0) However, slurmd logs the following error message 
for version 19.05.5 and 20.02.0 when start a GUI application, for example:

     [2020-03-11T11:42:55.673] [7.extern] error: _x11_socket_read: 
slurm_open_msg_conn: Connection refused

I checked $DISPLAY and that looked fine.

* (19.05.5 / 20.02.0) Here are some slurmd logs when start a Slurm 
session using "srun --x11 ..." :

     [2020-03-11T11:42:33.624] _run_prolog: run job script took usec=2200
     [2020-03-11T11:42:33.624] _run_prolog: prolog with lock for job 7 
ran for 0 seconds
     [2020-03-11T11:42:33.656] [7.extern] X11 forwarding established on 
DISPLAY=COMPUTE-NODE:97.0
     [2020-03-11T11:42:33.731] launch task 7.0 request from UID:3000 
GID:9001 HOST:HEAD-NODE PORT:25288
     [2020-03-11T11:42:33.732] lllp_distribution jobid [7] implicit auto 
binding: threads, dist 8192
     [2020-03-11T11:42:33.732] _task_layout_lllp_cyclic
     [2020-03-11T11:42:33.732] _lllp_generate_cpu_bind jobid [7]: 
mask_cpu, 0x00010001
     [2020-03-11T11:42:33.793] [7.0] in _window_manager
     [2020-03-11T11:42:33.798] [7.0] task_p_pre_launch: Using 
sched_affinity for tasks

* (18.08.7) Even though version 18.08.7 works, it always gives the 
following error message when start a GUI application, for example:

     [2020-03-11T13:54:21.112] [2.extern] error: _handle_channel: client 
disconnected
     [2020-03-11T13:54:21.113] [2.extern] error: _handle_channel: 
exiting thread

Any suggestions what could be done to get it work for version 19.05.5 / 
20.02.0 or what could be the cause of the problem?

Thank you.

Best regards,

Weijun