-
Notifications
You must be signed in to change notification settings - Fork 8
can't create report binding error #12
Comments
Hello, |
Thank you for looking into this. Please finde the config.yaml, cluster_config.yaml and log file in the attached zip archive. I am curious what you will find! |
Here a small update: I have tried installing culebrONT in different conda envs, thinking perhaps the error has something to do with the requirements. But so far not luck. Here is the environment.yml file of one of the environments I have tried out:
I has graphviz=2.50.0, python=3.11.5 and r-base=4.2.0. I usually create a conda environment first, activate it, install R, python and graphviz and then continue with the pip installation of culebrONT. Besides the run_report, my culebrONT pipeline also fails at the steps run_racon_version and blobtools. Do any of these rules have anything in common? Its just a pity that snakemake does not provide any insight as to what the error might be. This may be related to the particular snakemake version (see similar case here snakemake/snakemake#1698). I also noticed that the jobs fail, before they are submitted to slurm. Not sure, how to proceed. Perhaps a successful installation on my system is not possible. =( |
Hello ! thanks for files
Could you confirm us you are installing culebront in a slurm cluster? |
Yes, I want to use it on our institutional HPC cluster. I installed it with
But I am not sure, what this will achieve, when the log files are just empy, id doesnt matter where we save them? |
hello, the problem with report rule is because snakemake "script" key (into the report rule) sent stdout to the terminal and cannot be redirected to a log file. |
Hello, Have you file call Snakemake run Rmarkdown but can't redirect log on file. So normally this is write on terminal, on cluster this is write on slurm log. for exemple I have error with pandoc:
|
Not sure I understand. I need to wrap the culebrONT command in its own sbatch script? I previously just entered it on the command line. The pipeline itself then submits individual rules / jobs to slurm. Is that not correct?
|
@julieaorjuela Alright, I am not understanding it fully, but now I have run the
The file PS: So I am guessing this is a problem with how I use singularity (similar problem here: apptainer/singularity#5962). I have not used singulariy before and I am new to this particular HPC. So I guess its not surprising I dont have a user id. I just dont understand why the other rules, which also depend on software inside the singularity container, such as the fly and canu assemblies finish just fine? |
Can you send me the files:
|
@sravel Thanks for chipping in! I do have the described file. However it only contains:
I guess, it is because the pipeline already failed at the racon step and did not run all the way to the run_report rule. |
yes, in culebront command line you can also use snakemake parametters as --rerun-incomplete --nolock etc... first, please launch culebront with unlock param :
and add this params --rerun-incomplete --nolock to avoid this problem again such as:
and relaunch again. Julie |
and ... I wonder if report problem is not because yous singularity bind is strange |
I am happy to report, that I got the report! :)
In the end, I think the main culprit was that I did not define the singularity bind well. I hadn't used it before, and clearly specified it wrong (even though curiously it worked for some rules). PS: One more simple question. I can display the report just fine on the cluster. However, to share it with my team I need to pull it onto another fileserver. I tried to copy the complete FINAL_REPORT directory, but then the connection between the blobtools and assemblytics plots breaks. |
I am successfully running the test data set, but the pipeline stops at Error in rule run_report.
I am executing the command
culebrONT run_cluster --config /lustre/projects/dazzlerAssembly/test_culebrONT/data_test_config.yaml --singularity-args "--bind /lustre/projects/dazzlerAssembly"
This is the error I get
Unfortunately, there is no hint as to what the problem may be and both
/lustre/projects/dazzlerAssembly/test_culebrONT/culebrONT_OUTPUT/FINAL_REPORT/LOGS/REPORT.o
and/lustre/projects/dazzlerAssembly/test_culebrONT/culebrONT_OUTPUT/FINAL_REPORT/LOGS/REPORT.e
are empty.There is also no more information found in the
/lustre/projects/dazzlerAssembly/test_culebrONT/.snakemake/log/2023-09-14T152430.940894.snakemake.log
file.What could it be? I am guessing there might be a problem with required R packages or similar. But strictly this should not happen, as I am using the singularity container?
Do you have any ideas how I could debug this problem? I would really love to see the report!
The text was updated successfully, but these errors were encountered: