conda permission errors

[jsignell]

Originally posted as appveyor/ci#2822

Current Behavior

In Appveyor, using conda 4.6.1 to install any libraries which attempt to write an entrypoint script into the miniconda env directory raises Permission Denied errors and causes the build to fail. An example is this build on the public hvPlot repo:

ERROR conda.core.link:_execute(543): An error occurred while installing package 'conda-forge::ipykernel-5.1.0-py36h39e3cac_1001'.
PermissionError(13, 'Permission denied')
Attempting to roll back.
Rolling back transaction: ...working... done
[Errno 13] Permission denied: 'C:\\Miniconda36-x64\\envs\\earthml\\Scripts\\jupyter-kernel.exe'

Steps to Reproduce

conda install jupyter

Expected Behavior

On conda 4.5.4 the install works as expected

Environment Information

`conda info`

     active environment : None
       user config file : C:\Users\appveyor\.condarc
 populated config files : C:\Users\appveyor\.condarc
          conda version : 4.6.1
    conda-build version : not installed
         python version : 3.6.5.final.0
       base environment : C:\Miniconda36-x64  (writable)
           channel URLs : https://repo.anaconda.com/pkgs/main/win-64
                          https://repo.anaconda.com/pkgs/main/noarch
                          https://repo.anaconda.com/pkgs/free/win-64
                          https://repo.anaconda.com/pkgs/free/noarch
                          https://repo.anaconda.com/pkgs/r/win-64
                          https://repo.anaconda.com/pkgs/r/noarch
                          https://repo.anaconda.com/pkgs/msys2/win-64
                          https://repo.anaconda.com/pkgs/msys2/noarch
          package cache : C:\Miniconda36-x64\pkgs
                          C:\Users\appveyor\.conda\pkgs
                          C:\Users\appveyor\AppData\Local\conda\conda\pkgs
       envs directories : C:\Miniconda36-x64\envs
                          C:\Users\appveyor\.conda\envs
                          C:\Users\appveyor\AppData\Local\conda\conda\envs
               platform : win-64
             user-agent : conda/4.6.1 requests/2.18.4 CPython/3.6.5 Windows/2012ServerR2 Windows/6.3.9600
          administrator : True
             netrc file : None
           offline mode : False

`conda config --show-sources`

`conda list --show-channel-urls`

[msarahan]

Of course, I can't reproduce this locally. I think it might come down to an "all-users" vs "just-me" install. As a hack you could try adding an empty .nonadmin file to the C:\Miniconda36-x64 folder.

Sorry, but it might be a while before we can get around to debugging this. If you're stuck, I recommend just pinning to the older conda version for now.

[philippjfr]

As far as I can tell this is affecting all appveyor miniconda builds which install packages with entrypoint scripts, and presumably also a sizable portion of regular Windows users. Would be great to get a fix for this soon so we can start enjoying the speed improvements in latest conda 😄

I'll try reproducing it locally on my windows machine and dig a bit.

[kalefranz]

There's something that doesn't quite line up here. How is it that the package is conda-forge::ipykernel-5.1.0-py36h39e3cac_1001, but there's no mention of the conda-forge channel anywhere else in the report?

[kalefranz]

I see. That actual command according to https://ci.appveyor.com/project/pyviz/hvplot/builds/21921315/job/yr6fahphkmxw1cim#L720 was

conda install -y  -c pyviz/label/dev -c conda-forge "bokeh >=1.0.0" "holoviews >=1.11.0" "pandas<0.24.0" "coveralls" "nose" "flake8" "parameterized" "pytest" "nbsmoke >=0.2.0"

[jsignell]

Ah right - forgot to mention the channel info - like I said this was on appveyor so hard to debug.

[kalefranz]

My guess is the environment creation command was

conda create -y -n test -c pyviz/label/dev -c conda-forge "param>=1.7.0" "pyct>=0.4.4" "setuptools>=30.3.0"

So the two-step recreation process would be

conda create -y -n test -c pyviz/label/dev -c conda-forge "param>=1.7.0" "pyct>=0.4.4" "setuptools>=30.3.0"
conda install -y -n test -c pyviz/label/dev -c conda-forge "bokeh >=1.0.0" "holoviews >=1.11.0" "pandas<0.24.0" "coveralls" "nose" "flake8" "parameterized" "pytest" "nbsmoke >=0.2.0"

but it would be helpful to have that confirmed.

[msarahan]

There is a really neat trick to debug things on appveyor: https://www.appveyor.com/docs/how-to/rdp-to-build-worker/

[soapy1]

I was able to reproduce this on a windows machine running a non-admin console with the command

$conda install -y  -c pyviz/label/dev -c conda-forge bokeh holoviews pandas coveralls nose 
flake8 parameterized pytest nbsmoke

[msarahan]

I suspect that this has something to do with the entry point exe already being present in the package, then conda barfing when trying to overwrite it with a generated entry point exe. If that's accurate, the fix should be reasonable.

[kalefranz]

Using `-vv` should show you the full stack trace and you’ll probably see right where the problem is.

…

Sent from my iPhone

On Jan 28, 2019, at 5:13 PM, Mike Sarahan ***@***.***> wrote: I suspect that this has something to do with the entry point exe already being present in the package, then conda barfing when trying to overwrite it with a generated entry point exe. If that's accurate, the fix should be reasonable. — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.

[kalefranz]

Actually you only need one `-v`.

…

Sent from my iPhone

On Jan 28, 2019, at 5:13 PM, Mike Sarahan ***@***.***> wrote: I suspect that this has something to do with the entry point exe already being present in the package, then conda barfing when trying to overwrite it with a generated entry point exe. If that's accurate, the fix should be reasonable. — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.

[soapy1]

Looks like this is fixed on master, it should be resolved in the next release of conda

[ocefpaf]

I just tested conda 4.6.2 from the canary channel and it seems that the problem persists:

https://ci.appveyor.com/project/ocefpaf/notebooks-demos/builds/21998415

It may be specific to my config there. I'll try to isolate the problem and create some steps to reproduce it.

[msarahan]

This is so dumb, I had to laugh:

C:\Miniconda36-x64\envs\IOOS\Scripts\pygmentize.exe.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash - Access is denied.
C:\Miniconda36-x64\envs\IOOS\Scripts\tabulate.exe.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash.trash - Access is denied.

Clearly, needs improvement. I think we need to look at exactly what that post-link is failing on:

LinkError: post-link script failed for package conda-forge::nb_conda_kernels-2.2.0-py37_1000

The traceback indicates that it is related to AnacondaRecipes/nb_conda_kernels-feedstock#1

[dhirschfeld]

ping @SylvainCorlay - this might be of interest.

[SylvainCorlay]

Thanks @dhirschfeld. This appears to be it!

[github-actions]

Hi there, thank you for your contribution to Conda!

This issue has been automatically marked as stale because it has not had recent activity. It will be closed automatically if no further activity occurs.

If you would like this issue to remain open please:

1. Verify that you can still reproduce the issue in the latest version of Conda

2. Comment that the issue is still reproducible and include:

• What version of Conda you reproduced the issue on
• What OS and version you reproduced the issue on
• What steps you followed to reproduce the issue

3. It would also be helpful to have the output of the following commands available:

• conda info
• conda config --show-sources
• conda list --show-channel-urls

NOTE: If this issue was closed prematurely, please leave a comment and we will gladly reopen the issue.

In case this issue was originally about a project that is covered by the Anaconda issue tracker (e.g. Anaconda, Miniconda, packages built by Anaconda, Inc. like Anaconda Navigator etc), please reopen the issue there again.

Thanks!