conda install -Killed

Question

[marco.barr]

Hello everyone, surely this is something that has already been posted several times, but reading around on various topics, I can't find a definitive solution. How can I solve the issue that occurs when installing packages via conda? For example, I was trying this and it gives me this error that repeats with all the packages I want to install.

conda install -c nanoporetech -c conda-forge -c bioconda "nanoporetech::pychopper"         
      Collecting package metadata (current_repodata.json): done
      Solving environment: failed with initial frozen solve. Retrying with flexible solve.
      Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source.
      Collecting package metadata (repodata.json): - Killed

What can I do? Thanks to everyone for the support. (conda version: 22.9.0)

Answer 1

[i.sudbery]

Hi Marco. This isn't really specifically a bioinformatics problem, so its not really on topic here, but the message you are getting means that you ran out of memory because the solve was too complex. This usually happens when you have things installed that are not straight-forwardly compatible with each. conda was not designed to install everything into the same environment, but to make seperate environments for different tools/analyses. Try creating a new environment and installing the package in there.

[marco.barr]

Thanks so much Ian for your reply, the command I wrote above was already run in a conda environment created specifically for pychopper. My RAM is limited for the moment at 8 GB I don't know what to do

[dthorbur]

You might have better luck creating environments with mamba as it appears to be more memory efficient than conda. See discussion here. Another alternative is using an existing docker container, see list here. Though, if your machine is limited to 8Gb, I suspect any analysis with pychopper itself would run out of memory unless you ran it on very small batches.

[marco.barr]

Thanks, I managed to install mamba and with it also pychopper. I'll try it out