Difficulty Setting Up Seqera Tower Agent on in-house HPC Without Job Scheduler

Hello Seqera Community,

I am trying to set up the Seqera Tower Agent on our in-house HPC cluster as part of the Academic Program, but I’ve been facing several challenges and would greatly appreciate your guidance.

Environment details:

  • Our lab does not have access to AWS or GCP compute resources.

  • The in-house HPC is only accessible through university Wi-Fi or provided VPN (off-campus).

  • Strict network restrictions does not allow outbound port access or external command execution.

  • No job scheduler (e.g., SLURM, PBS, Altair) is available on the HPC.

Issue:
Since the Seqera Cloud Platform requires a job scheduler for integration, I have been unable to set up the Tower Agent successfully. I tried different workarounds and followed the documentation closely, but none of the attempts worked due to the above limitations.

Questions:

  1. Is there a way to configure Tower Agent in an HPC environment that does not have a job scheduler?

  2. Are there recommended lightweight schedulers or alternatives that I could consider installing to enable compatibility?

  3. Has anyone faced a similar restricted HPC setup and found a workaround?

Any advice, best practices, or references to relevant documentation would be very helpful.

Thank you in advance!

1 Like

Hey Muneed!

  1. Currently, there is not a way to launch from platform without a job scheduler.
  2. For a lightweight scheduler I’d recommend HyperQueue.
  3. There’s a feature request for this here, feel free to upvote it if it’s something you’d like to see!

You can launch your jobs with Nextflow run local or hyperqueue and use the -with-tower flag to enable monitoring in Seqera Platform!

```
-with-tower (https://api.cloud.seqera.io)

Monitor workflow execution with Seqera Platform (formerly Tower Cloud).
```

Thank you Edmund. Your response was quite helpful :grinning_face: