Skip to content

Resuming old jobs #2

Open
Open
@alexanderwerning

Description

@alexanderwerning

I ran autoexperiment, but my jobs crashed due to an error, so I stopped the autoexperiment process. I tried to restart them using autoexperiment on the next day, but I got these messages for each one:

Resume <exp> from job id: <job id>                                                                                                                                                                    
Current job id for <exp>: <job id>                                                                                                                                                                    
slurm_load_jobs error: Invalid job id specified                                    
Command 'squeue -j <job id>' returned non-zero exit status 1.                                                      
Retrying again in 10 mins for <exp>...  

It seems that the job ids are no longer tracked by slurm in squeue. Should I always manually delete all the logs/ files?
Is this behaviour intended?

Thank you!

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions