download failures

[beckermr]

We've been seeing a lot of download failures that look a lot like this:

https://github.com/ArgonneCPAC/dsps/actions/runs/12892731978/job/35947589532?pr=94

[jezdez]

Traceback for posterity:

  
      Traceback (most recent call last):
        File "/home/runner/miniconda3/lib/python3.12/site-packages/conda/exception_handler.py", line 18, in __call__
          return func(*args, **kwargs)
                 ^^^^^^^^^^^^^^^^^^^^^
        File "/home/runner/miniconda3/lib/python3.12/site-packages/mamba/mamba.py", line 960, in exception_converter
          raise e
        File "/home/runner/miniconda3/lib/python3.12/site-packages/mamba/mamba.py", line 953, in exception_converter
          exit_code = _wrapped_main(*args, **kwargs)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
        File "/home/runner/miniconda3/lib/python3.12/site-packages/mamba/mamba.py", line 899, in _wrapped_main
          result = do_call(parsed_args, p)
                   ^^^^^^^^^^^^^^^^^^^^^^^
        File "/home/runner/miniconda3/lib/python3.12/site-packages/mamba/mamba.py", line 768, in do_call
          exit_code = create(args, parser)
                      ^^^^^^^^^^^^^^^^^^^^
        File "/home/runner/miniconda3/lib/python3.12/site-packages/mamba/mamba.py", line 603, in create
          return install(args, parser, "create")
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
        File "/home/runner/miniconda3/lib/python3.12/site-packages/mamba/mamba.py", line 559, in install
          transaction.fetch_extract_packages()
      RuntimeError: Found incorrect download: ca-certificates. Aborting
  

[jezdez]

Raised this internally

[jezdez]

@beckermr we couldn't reproduce this ourselves, running conda create -n test-ca-certificates --override-channels -c conda-forge ca-certificates=2024.12.14 --platform=linux-64 worked for us.

1. Was this transient and resolved again?
2. Co you have a reproducer that we can use to find the cause?

[beckermr]

Hey thanks and sorry. More failures in this run:

https://github.com/ArgonneCPAC/diffdesi/actions/runs/12984582535/job/36207753041

They are indeed transient and usually pass on retries.

I do not have a reproducer yet.

I expect that making a reproducer is going to be very difficult.

Is it possible for you all to correlate the timestamps of errors on the backend with the timestamp of the download errors? That might at least help us figure out if it is an anaconda issue or maybe something on the github side.