job submit limit, user's size and/or time limits when run the case.submit

[CGL]

Hi,everyone. I try to run CESM2.1.3. I got the sbatch error:
ERROR: Command: 'sbatch --time 0:20:00 -p cpu_parallel --dependency=afterok:7408958 case.st_archive --resubmit' failed with error 'sbatch: error: QOSMinCpuNotSatisfied
sbatch: error: Batch job submission failed: Job violates accounting/QOS policy (job submit limit, user's size and/or time limits)' from dir '/data/sxh/CESM2/CESM/cgl/scratch/test2'

It seems like the case.st_archive limited the progress. Where can i change the limited size or set it?

 

[CGL]

I added the cesm.log report: SIGSEGV,segmentation fault occurred.

 

[nusbaume]

Hi CGL,

It looks like you have two different issues listed on this thread. For your first issue you can change the case.st_archive program’s queue and wallclock time by running the following commands in your case directory:

Job queue:

./xmlchange --subgroup case.st_archive --id JOB_QUEUE --val <value>

Wallclock time:

./xmlchange --subgroup case.st_archive --id JOB_WALLCLOCK_TIME --val <value>

Where <value> is whatever job queue name or wallclock time length you want.

In terms of your second issue, it looks like the error is happening in the atmosphere model (CAM). I first would check whether there is a useful error message in the atm.log. Otherwise I would try running with debugging on, which can be done by doing the following:

./xmlchange –id DEBUG –val TRUE

And then re-building and re-running the model, which should then hopefully give you a more specific error message.

Hope that helps, and have a great day!

Jesse

 

[CGL]

Hi CGL,

It looks like you have two different issues listed on this thread. For your first issue you can change the case.st_archive program’s queue and wallclock time by running the following commands in your case directory:

Job queue:


Wallclock time:


Where <value> is whatever job queue name or wallclock time length you want.

In terms of your second issue, it looks like the error is happening in the atmosphere model (CAM). I first would check whether there is a useful error message in the atm.log. Otherwise I would try running with debugging on, which can be done by doing the following:



And then re-building and re-running the model, which should then hopefully give you a more specific error message.

Hope that helps, and have a great day!

Jesse

it seems like the xmlchange do not have the DEBUG function.View attachment 3527

 

[CGL]

Hi CGL,

It looks like you have two different issues listed on this thread. For your first issue you can change the case.st_archive program’s queue and wallclock time by running the following commands in your case directory:

Job queue:


Wallclock time:


Where <value> is whatever job queue name or wallclock time length you want.

In terms of your second issue, it looks like the error is happening in the atmosphere model (CAM). I first would check whether there is a useful error message in the atm.log. Otherwise I would try running with debugging on, which can be done by doing the following:



And then re-building and re-running the model, which should then hopefully give you a more specific error message.

Hope that helps, and have a great day!

Jesse

I think the first question related to the second. Maybe some system limit the progress or other mistake. I will check this. Thanks for your reply:)

 

[ganbaranaito]

I think the first question related to the second. Maybe some system limit the progress or other mistake. I will check this. Thanks for your reply:)

Hello, have you solved the second issue? I have the same problem.

 

[CGL]

Hello, have you solved the second issue? I have the same problem.

I changed a cluster to run the model. It worked. I think the reason is the process or enviroment get wrong configuration when you try a parallel computing. You should contact with your cluster managers.

 

[ganbaranaito]

I changed a cluster to run the model. It worked. I think the reason is the process or enviroment get wrong configuration when you try a parallel computing. You should contact with your cluster managers.

Thank you!