geodesymiami / rsmas_insar

RSMAS InSAR code
https://rsmas-insar.readthedocs.io/
GNU General Public License v3.0
59 stars 23 forks source link

—rapid option for submit_jobs.bash plus parameter summary #476

Closed falkamelung closed 3 years ago

falkamelung commented 3 years ago
Ovec8hkin commented 3 years ago
  • Currently submit_jobs.bash sleeps 60 seconds and 300 seconds (5 min) before checking the state of a job (using sacct) and before trying to submit a new job when the MAX_JOB_LINITS_PER QUEUE (1, 25 for skx-dev, skx-normal) is reached, respectively. This work fine most of the time, but when running a small test workflow on skx-dev it will be better to have short wait times (20 seconds and 60 seconds, respectively). Can we swicth from hard-wired numbers to variables? The defaults would be 60,300 but an —rapid option would set to 20,60 ((submit_jobs.bash $PWD —start 7 —rapid )

Just pushed this.

  • currently it is a bit messy which are the parameters of submit_jobs.bash and where they are set (I have seen different tasks limit of (3000) and skx-normal (1000 or 1500), I think). Could you add some explanation on this? A possible location is the —help message. Also please explain in the script where the parameters are defined. in the queues.cfg and overwritten if an environment variable is declared? I forgot what we decided to do. I admit that I said many times 'use en environment variable'.

Not sure what you're referring too here.