usegalaxy-no / galaxyadmin

A repository for managing the work of the usegalaxy.no GalaxyAdmin team
0 stars 0 forks source link

Tool job errors from weeks 6 and 7, 2022 #65

Closed kjetilkl closed 2 years ago

kjetilkl commented 2 years ago

Here is a summary of the tool problems found between February 4th and 20th that are not due to user mistakes. The numbers in brackets behind the tool names refer to the job ID of a failed job execution. (There can be more failed jobs for the same tool as well)

The following tool jobs failed because Galaxy was unable to build the Singularity containers at the time of execution due to "no space left on device". However, the containers could be successfully rebuilt by re-running the tools at a later time, so they all appear to work now.

The following tools had existing containers, but these were missing some required libraries, so I installed new containers from BioContainers. This fixed "samtools_sort" (Sort BAM dataset), but "bwa_mem" now fails with a different error. Both tools have newer versions installed on UseGalaxy.no that work just fine.

The "fastq-join" tool required the package "ea-utils", which was only available as an old-style toolshed package and was not available in Conda (for the required version). To fix this I had to manually build a new Docker image for this package based on the Dockerfile available from the "ea-utils" GitHub repository and then convert this to a Singularity image.

The next two tools fail with the error "command not found". I have installed containers for both of these from BioContainers, and the containers appear to work fine when the tool command is run manually. However, they still fail when the tools are run in Galaxy, and I do not know why. Requesting assistance!

In addition, the RNA Star tool apparently failed because there wasn't enough memory to run the job [144550] (February 16). We should probably look into that.

The "Get files from NeLS storage" tool (aka nels_file_browser) failed with an error which I suspect may be a user error, but it is cryptic enough for me to start doubting [145035] (February 17)

pysam.utils.SamtoolsError: 'samtools returned with error 1: stdout=, stderr=samtools index: failed to create index for "/data/part0/000/912/dataset_912150.dat": Inappropriate ioctl for device\n'

And the "omics@icloud.com" user has been trying to run the "interactive_tool_jupyter_notebook" tool, which we don't really support yet (so we should probably remove it until we do)

kjetilkl commented 2 years ago

I accidentally stumbled upon the reason why the tools circos and goenrichment did not work, even though their containers seemed to be correct. It turns out that these two tools (and three others as well) were configured in tool_destinations.yaml to use the "slurm" job runner rather than the default "slurm_singularity" job runner. Because of this, the tools were not executed inside containers.