Bug 5692 - X11 forwarding error: _get_home: getpwuid_r(33560764):No error
Summary: X11 forwarding error: _get_home: getpwuid_r(33560764):No error
Status: RESOLVED INVALID
Alias: None
Product: Slurm
Classification: Unclassified
Component: Scheduling (show other bugs)
Version: 17.11.9
Hardware: Linux Linux
: --- 6 - No support contract
Assignee: Jacob Jenson
QA Contact:
URL:
Depends on:
Blocks:
 
Reported: 2018-09-09 15:31 MDT by Jack Duvall
Modified: 2018-09-09 16:46 MDT (History)
0 users

See Also:
Site: -Other-
Alineos Sites: ---
Atos/Eviden Sites: ---
Confidential Site: ---
Coreweave sites: ---
Cray Sites: ---
DS9 clusters: ---
HPCnow Sites: ---
HPE Sites: ---
IBM Sites: ---
NOAA SIte: ---
OCF Sites: ---
Recursion Pharma Sites: ---
SFW Sites: ---
SNIC sites: ---
Linux Distro: ---
Machine Name:
CLE Version:
Version Fixed:
Target Release: ---
DevPrio: ---
Emory-Cloud Sites: ---

Attachments
Verbose error log for issue (11.62 KB, text/plain)
2018-09-09 15:31 MDT, Jack Duvall 		Details
slurm.conf (4.00 KB, text/plain)
2018-09-09 15:32 MDT, Jack Duvall 		Details
Verbose slurmd.log, 18.08 version (12.05 KB, text/plain)
2018-09-09 16:46 MDT, Jack Duvall 		Details
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Jack Duvall 2018-09-09 15:31:08 MDT 
Created attachment 7790 [details]
Verbose error log for issue

Hello,

I'm having an issue setting up X11 forwarding. Everything was compiled with the necessary libssh2 and libssh2-devel libs, X11 forwarding enabled using the PrologFlags=x11 switch, and srun itself doesn't complain, yet any graphical programs throw an error about not being able to open the display.

Digging further in the logs for the node, I found this (debug2 level logging file attached if you want to see the full output):

> [2018-09-09T16:52:20.042] _run_prolog: run job script took usec=9
> [2018-09-09T16:52:20.042] _run_prolog: prolog with lock for job 2064 ran for 0 seconds
> [2018-09-09T16:52:20.063] [2064.extern] error: _get_home: getpwuid_r(33560764):No error
> [2018-09-09T16:52:20.063] [2064.extern] error: could not find HOME in environment
> [2018-09-09T16:52:20.064] [2064.extern] error: x11 port forwarding setup failed
> [2018-09-09T16:52:20.065] [2064.extern] error: _spawn_job_container: failed retrieving x11 display value: No error
> [2018-09-09T16:52:20.095] [2064.extern] done with job
> [2018-09-09T16:52:20.157] launch task 2064.0 request from 33560764.2019@198.38.16.106 (port 181)
> [2018-09-09T16:52:20.159] error: could not get x11 forwarding display for job 2064 step 0, x11 forwarding disabled
> [2018-09-09T16:52:20.216] [2064.0] done with job

The part about not being able to find $HOME is strange, because the output of `srun -n1 -p compute2 -w borgw201 --x11 /usr/bin/env | grep HOME`, the same command that produced the error above, says $HOME is defined.

One thing of note about my setup: Unlike in past errors, users are not allowed to ssh to compute nodes. All authentication is done with kerberos, so ssh keys don't exist. User home directories are on a shared NFS drive too. If this error is just a symptom of not having ssh key-based authentication and X11 forwarding isn't supported under any other setup, that would be nice to know for certain.

I haven't had time yet, but will try the latest version of slurm (18.08.0) and close the issue if updating fixes it.

If not, I will attach my slurm.conf next, and can provide any other files requested.

Thanks,
-Jack Duvall
Comment 1 Jack Duvall 2018-09-09 15:32:45 MDT 
Created attachment 7791 [details]
slurm.conf
Comment 2 Jack Duvall 2018-09-09 16:46:04 MDT 
Update: 18.08.0 does not fix this issue for me. Nothing seemed to change drastically in the logs either.

New abbreviated log:

> [2018-09-09T18:40:31.983] _run_prolog: run job script took usec=169
> [2018-09-09T18:40:31.984] _run_prolog: prolog with lock for job 2067 ran for 0 seconds
> [2018-09-09T18:40:32.046] [2067.extern] error: _get_home: getpwuid_r(33560764):No error
> [2018-09-09T18:40:32.047] [2067.extern] error: could not find HOME in environment
> [2018-09-09T18:40:32.047] [2067.extern] error: x11 port forwarding setup failed
> [2018-09-09T18:40:32.065] [2067.extern] error: _spawn_job_container: failed retrieving x11 display value: No error
> [2018-09-09T18:40:32.065] [2067.extern] error: _spawn_job_container: failed retrieving x11 authority value: No error
> [2018-09-09T18:40:32.085] [2067.extern] done with job
> [2018-09-09T18:40:32.102] launch task 2067.0 request from UID:33560764 GID:2019 HOST:198.38.16.106 PORT:17035
> [2018-09-09T18:40:32.104] error: could not get x11 forwarding display for job 2067 step 0, x11 forwarding disabled
> [2018-09-09T18:40:33.843] [2067.0] done with job

debug2 version of log attached.
Comment 3 Jack Duvall 2018-09-09 16:46:36 MDT 
Created attachment 7792 [details]
Verbose slurmd.log, 18.08 version