Juicer job failing due to Java memory allocation overriding SLURM limits

[Gabriela Jijon]

Hi all, 

My juicer job is failing because Juicer's scripts (juicer.sh) seem to override my SLURM resource limits, causing Java to request more memory than I have been allocated (e.g., requesting ~150 GB). This results in Java steps crashing. 

Is there a recommended way to force Juicer/Java to respect SLURM memory limits on this cluster? 

Thank you in advance for your help. 

Best, 

Gabi