PBS job is detected, but nodes do not power on

[daantreurniet]

Dear team,

I am trying to set up CLUES for a HPC using Torque with PBS as queue manager. After some configurating, I managed to start the CLUES server succesfully with the PowerOn_Requests scheduler enabled. I can also see that the queue is detected, because when I submit a new job, it is shown in the CLUES server output. However, after the job is detected, no nodes are powered on to start working. All nodes stay powered off and the job remains in the queue, waiting for resources. I tested the poweron and poweroff commands via the clues CLI (I'm using IPMI to command the nodes) and that is working properly.

Any idea what the problem could be? I hope this project is still being monitored, as it would be very useful!

Best regards

[micafer]

Dear @daantreurniet,

Could you attach the clues logs to check what can be the problem?

Best regards.

[daantreurniet]

When starting cluesserver manually without specifying a Log file, I get the following output. It keeps going with messages similar to the last 6 in the log file. log.txt