I have sent an error report 11ac94870d0bb33ad2018af257e57a4a from within usegalaxy.eu.
Am I supposed to post this request here again, or has this technical support request already registered at usegalaxy.eu? The error messages at galaxy were:
Command Line
q2galaxy run feature_classifier fit_classifier_naive_bayes ‘/data/jwd06/main/097/093/97093712/configs/tmpoe_i4gcr’
stderr
docker: Error response from daemon: error while creating mount source path ‘/cvmfs/data.galaxyproject.org’: mkdir /cvmfs/data.galaxyproject.org: file exists.
You can always submit a bug report, and the server administrators do appreciate that! They review for server issues as the priority. If it is not clear that the error is technical, then it is unlikely that you’ll get a custom reply.
Instead, or in addition to, you can ask here and we’ll troubleshoot to get the error triaged to the right places. Knowing where in the chain the error is coming from can be quite complicated!
Your error appears to be from the job environment failing to build. The UseGalaxy.eu had a large update over the last several weeks specifically related to the protocol recipes for these job environments. Not the actual dependencies, more about how they are assembled at runtime. Capturing and reporting back what may still be unresolved is important and we can do that here! Example of similar tool error earlier today.
How you can help us to investigate:
Share link to a small history with the error. If the inputs are intact, that helps! FAQ: Sharing your History
Or, sometimes the job details page is enough. Make sure we can see the tool version, the paramaters used, expand inputs peek views, and all logs, including the command string and both stderr and stdout. FAQ: Troubleshooting errors
For this tool, the full tool name with the version is probably enough! I may be able to reproduce with the defaults and test data. My guess is that the Qimme2 docker itself is conflicting with some local settings on the server, or maybe the problem is happening for everyone due to a version update.
Does this look correct? I’ll start testing but please confirm.
qiime2 feature-classifier fit-classifier-naive-bayes Train the naive_bayes classifier (Galaxy Version 2026.1.0+q2galaxy.2026.1.0)
Then, if you want to explore a workaround, sometimes toggling into the prior tool version is enough (if this was due to an update). I haven’t tested this yet for this specific issue, and will, but you can also try since it might get you going sooner! Sometimes going back once is enough, sometimes more.
Thanks for reporting the issue and more soon! I’ll watch for your confirmation that I guessed the right tool/version combination.
This is almost correct, but I used qiime 2025.4 instead of the most recent 2026.1 version. This is due to the qiime2 artifact architecture which is not backward-compatible (a qiime2 classifier created with 2025.10 or 2026.1 cannot be used e.g. with 2024.10).
Would you like to use my seq and tax qza files for this run? I could share these documents. Please keep me updated on the outcome.
I have also created an issue report as requested under point 2 above (I did this already last week). I hope the administrators will be able to see it.
Yes, seeing the exact run would be very helpful. If you want to share back your history that would be fast and exact! You can copy just the inputs/outputs of the failed run into a new history to limit the content FAQ: Sharing your History
I was trying to build up the example yesterday and was sidelined. The easy case is “any data” will trigger this but if it turns out to be more complicated, any of other details may be involved.
after receiving a response from administrators that the issue should be fixed now, I restarted the exact run again , but this job is still running. I will let you know about the outcome.
