Cryosparc disconnect and job stop to run

Hello,
In our team, we are currently working on a Linux workstation with Cryosparc V4.4.1. We have had problems since the beginning. When we run jobs, Cryosparc sometimes disconnects but seems to be still running. However, the progress of the job is stopped (without an error message). When this happens, we have to delete the “.sock” file, on TMP files, in order to restart and reconnect to Cryosparc.
On rare occasions, we don’t have this problem during the week; however, most of the time, these stoppages occur several times during the day, at any given time.
These interruptions occur on any job (hetero/homo/NU refinement, patch motion correction, AB initio correction, 2D classifications, etc.). We are convinced that this is not a network problem, as we have a second Linux workstation with cryosparcm V4.2.1 connected on the same network, and it does all the jobs without this problem.
Is this due to a cryosparcm version problem? Is there a way to fix it?

Thank you in advance for your reply.

Welcome to the forum @ClaraB.
Please can you

  1. show a screenshot of the UI when
  2. when CryoSPARC has apparently disconnected, collect and post the output of these commands
    ps -eo user:16,pid,ppid,start,rsz,vsz,command | grep cryosparc_
    free -h
    cat /sys/kernel/mm/transparent_hugepage/enabled
    ls -l /tmp/cryosparc*.sock /tmp/mong*.sock
    
  3. post the full URL, including the protocol (https or http) from the web browser’s address bar when connected to the CryoSPARC UI
  4. describe how you confirmed that CryoSPARC

You may want to avoid deleting CryoSPARC-related *.sock files
before having confirmed that all related CryoSPARC processes have exited (details).