I would like to be able to launch jobs directly on share workstations without having to install a scheduler. Nextflow should simply run the job as “local” instead of “slurm”, “pbs”, etc… This feature would be extremely useful as it would help managing those workstation (I would create a workspace for each one) without having to provide access to each user. It would work either with ssh or tw agent. I proposed this feature to seqera, but maybe there is some way to do it that I missed.
What happens if you set the executor to local in a config? I don’t think I’ve ever tried it. Might be that it works, and it doesn’t matter that you set the compute environment up as Slurm even though Slurm isn’t available.
AFAIK there is no official support for launching local executor from Seqera platform. See this issue for more details.
The only current solutions are:
- To install a single machine scheduler (using docker may not be that difficult)
- Using the local executor and creating some Slurm fake scripts. I also did some fake scripts to submit and monitor jobs to a HyperQueue scheduler, which is a very easy-to-install solution if you want to use just a few workstations.
These workarounds should work with the Agent and when directly using an SSH connection.
Thanks @jordeu , docker installation may be a good way to do this. Probably we should need a docker in docker image, and add aliasaes (alias sinfo=“docker exec -ti ${USER}_simulator sinfo” etc.). I’ll have a try.
For any future googlers: there is a feature request to track this here:
Please add your votes if this is important to you, and you’ll be able to track any progress there. Thanks!
1 Like
To follow up on @jordeu proposal, this is what we are doing now. We extended the docker slurm image with the tower agent, this is working fine.