Hello community,
I manage a NGS core lab and also teach a grad level intro to Bioinformatics course. We used to have a private instance of Galaxy on our servers but we had all sorts of permission issues. I recently was awarded an Xsede allocation and I have launched a pre-configured instance of Galaxy on Jetstream. Now I thought my issues with getting tools to work would be over since this pre-configured Galaxy has Conda, well I was wrong. Some of the tools that are pre-loaded don't even work. Specifically, I am trying to get Trimmomatic or Cutadapt to work but I get this error:
ERROR: A required extension module could not be imported because it is incompatible with your system. A quick fix is to recompile the extension modules with the following command:
/opt/galaxy/galaxy-app/database/jobs/000/14/conda-env/bin/python setup.py build_ext -i
See the documentation for alternative ways of installing the program.
The original error message follows.
Can anyone help? I tried to recomplile as suggested but there is no file called setup.py in the specified directory.
Is there additional context to the error message? This almost looks like a corrupt conda installation, although there is something else strange here, which is that Galaxy is creating a per-job Conda environment when you run the tool. That should only happen if your tools' dependencies are not "mulled" in to a single Conda environment.
You should be able to see what's going on with your Conda environments using the "Manage dependencies" interface under the Admin section of Galaxy.
How can I provide additional context? I was under the impression that this Galaxy would be completely pre-configured to mirror the public Galaxy instance. I have not changed anthing other than making myself an Admin in the galaxy.ini file. There is no "Manage dependencies" interface. Here is additional text from the error message:
The Jetstream image does attempt to mirror the same tools as usegalaxy.org but beyond that it's very different from usegalaxy.org in the way it is configured and tools are installed.
This situation looks pretty ugly. Part of the issue stems from the fact that Galaxy in Jetstream is currently version 17.01 (this is also why you don't have the manage dependencies interface), newer versions have resolved these issues.
What's happened here is that Galaxy is calling cutadapt which was installed via Conda, but then is resolving cutadapt's python dependencies from a different version that was installed via the old "tool shed dependency" method.
As a workaround, you may be able to simply move
/opt/galaxy/tool_dependencies/cutadapt/1.8
(and the/opt/galaxy/tool_dependencies/cutadapt/default
symlink if it exists) out of the way and try running the tool again.For a proper solution, I'll see if we can get the Galaxy version in the Jetstream image updated.
Hi Nate,
Thank you for all of your input. Can you please tell me if the Galaxy version in Jetstream has been updated. I am having a heck of a time installing any additional tools to my instance of Galaxy.