Skip to content

Latest commit

 

History

History
34 lines (27 loc) · 752 Bytes

faq.md

File metadata and controls

34 lines (27 loc) · 752 Bytes

FAQ

How to enforce local scheduling on a cluster?

Define a bash variable named JL_FORCE_LOCAL:

export JL_FORCE_LOCAL=true

This will cause jl to use the local scheduler even if an actual queueing system is available. It will indicate this by printing the message

jl status
.jobs: Using local scheduler
...

It's even possible to jump between schedulers:

jl submit "sleep 100"
jl kill

## continue locally
export JL_FORCE_LOCAL=true
jl resub --retry

This is especially handy to debug failing jobs.

To restore scheduler auto-detection, simply remove the variable with

unset JL_FORCE_LOCAL

How to use jl to monitor a large number of jobs?

See section Batch Submission in the user manual