Error in IMPR_RESU: unable to creae directory

jcsaaver

I was setting up a model for computation, and got an error with the command IMPR_RESU.
It appeared when, in Asterstudy, I clicked the 3 dots to create the result file.
The terminal shows this (some messages are repeated several times, I pasted them only once):

SalomeApp_Engine_i::Load() isMultiFile = 0
AsterStudy: sam.-29-02:16:19.083 Opening AsterStudy JSON file...
AsterStudy: sam.-29-02:16:19.088 Use version 'stable' (17.4.0)
AsterStudy: sam.-29-02:16:19.088 loading case 'CurrentCase' (1/1)...
AsterStudy: sam.-29-02:16:19.089 loading stages...
AsterStudy: sam.-29-02:16:19.276 case 'CurrentCase' loaded
AsterStudy: sam.-29-02:16:19.632 JSON file reloaded
AsterStudy: sam.-29-02:16:20.682 AsterStudy is activating...
AsterStudy: sam.-29-02:16:20.683 Creating workspace...
AsterStudy: sam.-29-02:16:20.732 Refreshing configuration on localhost...
AsterStudy: sam.-29-02:16:29.539 workspace is ready.
AsterStudy: sam.-29-02:16:31.260 AsterStudy activated

Error creating proxy: No se pudo conectar: No existe el fichero o el directorio (g-io-error-quark, 1)
(SALOME_Session_Server_No_Server:4046): dconf-CRITICAL **: 02:17:06.503: unable to create directory '/run/user/1000/dconf': Sistema de ficheros de sólo lectura.  dconf will not work properly.

XXX lineno: 486, opcode: 124
XXX lineno: 486, opcode: 150
XXX lineno: 35, opcode: 151
Error in sys.excepthook:
SystemError: unknown opcode

Original exception was:
SystemError: unknown opcode
XXX lineno: 486, opcode: 150
XXX lineno: 35, opcode: 151
Error in sys.excepthook:
SystemError: unknown opcode

Original exception was:
SystemError: unknown opcode
XXX lineno: 486, opcode: 124
XXX lineno: 35, opcode: 116

I suppose it tries to create temporal files somewhere it has no permission, but I could not find how to change thap place, wich appears to be /run/user/1000/dconf

Any ideas?
(I tried to upload my hdf file, but I couldn't)

mf

jcsaaver
Hello,

I am also getting this exact error in SM2025 on Ubuntu 24.04 on a machine without nvidia (software-rendering). The other machine with nvidia works without this error. If I remember correctly, IMPR_RESU in SM2023 also had a problem regarding this window (there's a message about this in this forum somewhere).
I start the container like this:

singularity run --app install salome_meca-lgpl-2025.1.0-1-20251026-scibian-12.sif --bind /tmp:/tmp --bind /tmp:/local00/tmp -B /media/drive2
./salome_meca-lgpl-2025.1.0-1-20251026-scibian-12

But when I start the container like this:

singularity exec \
--bind /tmp:/tmp \
--bind /tmp:/local00/tmp \
-B /media/drive2 \
salome_meca-lgpl-2025.1.0-1-20251026-scibian-12_CA175_mpi.sif \
/opt/salome-meca/2025/V2025.1.0_scibian_univ/prerequisites/Cea_archive-9130/salome

the error does not occur. This is also how I start it on the machine with nvidia. So it's not the --nv or software-rendering, it is the "singularity exec...".

Another observation is, that if I press the ... button, the folder that is open is always ../. So when my Study folder is in /home/user/testcase, then the ... button sends me to /home/user. Maybe that has something to do with it?

Mario.

basile-simvia

Can you tell me which version of Salome Meca you're using and how you launch Salome Meca from the terminal?

basile-simvia

I just published a comment in the dedicated youtube video to explain the problem and propose a fix. https://www.youtube.com/watch?v=WtiAcHkdiU4

mf

basile-simvia
Just for info what is installed on my desktop with nvidia:

$ apt list squashfuse fuse2fs gocryptfs
fuse2fs/noble-updates,now 1.47.0-2.4-exp1ubuntu4.1 amd64 [installed]
gocryptfs/noble-updates,noble-security 2.4.0-1ubuntu0.24.04.3 amd64
squashfuse/noble 0.5.0-2build1 amd64

jcsaaver

I found that in my case the problem was that I was running salome_meca without using the original script, as I had extracted the .sif container. So I had to modify one line of the script for it to run properly.
The solution is described here.