VFLP vs VFVS changed environment needs?

VFLP VFLP 1
#1

Hi,

I have VFVS up and running, and one major hurdle was configuring an LSF environment to get it running.

I’m trying to get VFLP going now, using the same environment, but it’s failing. I am having trouble determining why. I have vftools, OpenBabel, JAVA14, jchemsuite (and license file), and nailgun installed - hence I don’t think this has to do with any external prerequisites.

Is there some difference in the cluster environments needed to get VFLP vs VFVS running successfully?

Thanks,
Chris

#2

Hi Chris,

one issue could be jchemsuite. It depends on which version you have. ChemAxon recently remove a component in jchemsuite from the free academic license which VFLP uses. So I would recommend to look at the logfiles in the workflow/output-files/queues/… folders to find out the error.

Best,
Christoph

#3

Thanks for your response Christoph,

I’ve been trying to work this out but still running into issues. Despite have ligands set up in the correct hierarchy and the same environment as VFVS, the program doesn’t actually make it far enough to produce logfiles in workflow/output-files/queues/

Theoretically, the environment should be the same and once we input metatranches containing ligands (.smi format) and the external packages it should successfully run correct? There’s nothing else that should need to be changed in an LSF environment?

Thanks,
Chris

#4

Hi Chris,

How about logfiles in workflow/output-files/jobs/ ? If there are some, can you post one here in the forum?

With “the same”, to which other environment are you referring to?
Did you run one of the tutorials successfully?

Best,
Christoph

#5

Thanks Christoph,

I’ve run both VFVS tutorials successfully and also used VFVS for my own targets within an LSF cluster environment, so that’s why I think the environment and setup should be ok. I’ve got all the ‘packages’ within the tools subfolder, including a specific JChem license that includes JChem suite (not the just the academic version), and my ligands are assembled as a meta-tranche.

Attached is a sample workflow/output-files/jobs/ job file. Each job runs for ~2 min before stopping/failing. Is there anything in here that would help identify the problem for you?

Thanks,
Chris
ThomC.JobFile.rtf (41.2 KB)

#6

Hi Chris,

This is really unusual.

Running an interactive job, and then manually running the main job script from the tools folder with something like
bash ../workflow/job-files/main/1.job
might give more information.
Also, placing a line in one-step.sh file in the beginning with something like
echo "Entering one-step.sh"
might give us a clue if the script is entered at all.

Best,
Christoph