Failed to create symbolic link

Hey guys,

One of my jobline seems to be stuck in a loop and has been submitting very small jobs, of around 2 minutes for over 2 days continously, without being able to refill the ligands. It seems to be that the todo.all file in workflow/ligand-collections/todo/ is empty but ./vf_report.sh -c workflow shows that that just a couple of ligands left (5k out of 15M).

A todo.all.locked file has been created in /tools/helpers/ but when I want to inspect the file it says file doesnt exist.

Furthermore, every submitted job shows the folling error in the logfile:

  Preparing the to-do lists for jobline 5
Sun Aug  2 09:28:22 CEST 2020
prepare-todolists.sh: line 98: 10#locked: value too great for base (error token is "10#locked")
wc: todo.all.locked: No such file or directory
ln: failed to create symbolic link './todo.all.locked': File exists
The next todo list will be used ()
cp: omitting directory '../../workflow/ligand-collections/todo/'
There is no more ligand collection in the todo.all file. Stopping the refilling procedure.
Starting job step 1 on host

Hi Arthur,

Not sure what went wrong with the file. There never should be a todo.all.locked file in the /tools/helpers folder. Did you run some of the helper scripts manually? If not, maybe restarting the workflow would be best. If the problem repeats, then we should look more closely into it.

Best,
Christoph

Hello,

During the screening a job was held by slurm and I released it manually using scontrol release <job_id>. I continued the screening by ./vf_continue_jobline.sh script. After this I noticed that there is just a couple thousand of ligands left to be screened and they can’t be refilled due to the locked file.

I’ve cancelled the jobs using scancel and again used ./vf_continue_jobline.sh but the same problem is still ocurring.

The problem is that when I want to extract the docking poses with the head and
vfvs_pp_prepare_dockingposes.sh commands, I don’t get all my results since some collections are not finished e.g.

./vf_report.sh -c vs -d qvina02_rigid_receptor1 -n 10 shows a couple of different compounds which are not included when running the head command.