It4innovations / hyperqueue

Scheduler for sub-node tasks for HPC systems with batch scheduling
https://it4innovations.github.io/hyperqueue
MIT License
273 stars 21 forks source link

Support for Sun Grid Engine (SGE) #523

Open bguo068 opened 1 year ago

bguo068 commented 1 year ago

Hi, I am new to this project but feel very excited about the idea of sub-node tasks. I tried the hyperqueue on the clusters managed by Sun Grid Engine(SGE) in our institute. The manual worker deployment works fine. However, when I tried to use Automatic worker deployment (via PBS, I would imagine PBS and SGE work similarly).

hq alloc dry-run  pbs --time-limit 1h -- -P  xxxlab

it gave me the error:


Error: Received error: 

Could not submit allocation: qsub execution failed

Caused by:
    Exit code: 1
    Stderr: Unable to run job: job rejected: no project assigned to job.
    Exiting.
    Stdout: 
```.
Did I make some mistakes using cli arguments? Or SGE is not supported yet? 
Thanks!
Kobzol commented 1 year ago

Hi, there is currently no explicit support for SGE in the auto allocator. We don't have access to a SGE cluster at the moment, but I will take a look at how does SGE differ from PBS/Torque and try to implement support for it.