Cluster queuing for Import Movies jobs is not working under version 2.14.2 for us. When clicking on Queue, the resulting window does not show the Queue Type option but rather just the Lane option. Selecting the Lane option, which is what we do normally anyways, does not submit to the queue but rather runs the job on the master node. Other job types are not affected and are submitting to the queue as expected.
This is the expected behaviour for import jobs as they do not require any specific compute requirements. Is there a particular reason why you would like the import movies job to run on a cluster node?
No particular reason beyond wanting to manage our resources. Since these jobs fall outside of our scheduling system, they are not tracked. It just seemed strange to exclude one job type from being able to be queued.
Thanks for your help with this,
In order to queue master-only jobs to worker nodes (including cluster nodes):
Navigate to cryosparc2_master and add the following line to the end of config.sh:
Once you’ve saved config.sh, restart cryoSPARC:
Now, import jobs (volume, particles, templates, movies & micrographs) will run on the target of the user’s choice.
Please note that interactive jobs are required to be run off the master node, regardless of this variable.