Conda Install Fails

[jjoonathan]

conda install -c pyviz geoviews currently fails to Solve Environment inside a clean anaconda container launched with sudo docker run -it continuumio/anaconda3.

I've tried the pyviz channel, the default channel, and the conda-forge channel with the same results.

Conda suggests that the problem has to do with conflicts and that it is looking for incompatibilities, but it has been running since yesterday and only reports 90% progress. I'll update this post if it completes.

Found conflicts! Looking for incompatible packages. This can take several minutes. Press CTRL-C to abort. Examining conflict for ...

[jjoonathan]

Here is conda info inside the container:

sudo doc(base) jon@jon-Precision-7720:~/hyperspectral$ sudo docker run -it continaconda3 
(base) root@68c9563607c9:/# conda info

     active environment : base
    active env location : /opt/conda
            shell level : 1
       user config file : /root/.condarc
 populated config files : 
          conda version : 4.8.2
    conda-build version : 3.18.11
         python version : 3.7.6.final.0
       virtual packages : __glibc=2.28
       base environment : /opt/conda  (writable)
           channel URLs : https://repo.anaconda.com/pkgs/main/linux-64
                          https://repo.anaconda.com/pkgs/main/noarch
                          https://repo.anaconda.com/pkgs/r/linux-64
                          https://repo.anaconda.com/pkgs/r/noarch
          package cache : /opt/conda/pkgs
                          /root/.conda/pkgs
       envs directories : /opt/conda/envs
                          /root/.conda/envs
               platform : linux-64
             user-agent : conda/4.8.2 requests/2.22.0 CPython/3.7.6 Linux/5.3.0-42-generic debian/10 glibc/2.28
                UID:GID : 0:0
             netrc file : None
           offline mode : False

[jjoonathan]

conda install -c pyviz geoviews-core completes successfully. It's just the full geoviews that conda has issues with. I wish it didn't take so long (~an hour) to fail, so that I could iterate quicker, but that's how conda is.

Update: conda's dependency analysis finally finished after 18 hours! Unfortunately it does not look terribly helpful at first glance. Hopefully it's more helpful to one of you :)
conda_conflict_report.txt

[jjoonathan]

Ok, one more possibility eliminated. I don't think it's just the continuumio/anaconda3 docker image.

I manually started with sudo docker run -it ubuntu:latest, installed anaconda apt update; apt install curl; curl -O https://repo.anaconda.com/archive/Anaconda3-2020.02-Linux-x86_64.sh; bash ./Anaconda3-2020.02-Linux-x86_64.sh, and ran conda install -c pyviz geoviews . It failed to Solve Environment.

[jjoonathan]

I've started manually installing combinations of geoviews dependencies. Unfortunately, whether or not they Solve Environment seems to be order-dependent. In each of these cases, I start with a fresh container sudo docker run -it continuumio/anaconda3:

I can successfully install all of the following (copied from history, might contain a typo):

conda install -yc pyviz geoviews-core
conda install -yc pyviz bokeh
conda install -yc pyviz datashader
conda install -yc pyviz xarray
conda install -yc pyviz holoviews
conda install -yc pyviz jupyter
conda install -yc pyviz matplotlib
conda install -yc pyviz numpy pandas
conda install -yc pyviz param
conda install -yc pyviz netcdf4
conda install -yc pyviz python
conda install -yc pyviz scipy
conda install -yc pyviz cartopy
conda install -yc pyviz pyct
conda install -yc pyviz shapely

If I subsequently try either of the following, I get a failed Solving Environment:

conda install -c pyviz geopandas
conda install -c pyviz gdal

If I try to install the "working" set of packages all at once, I get a failed Solving Environment:
conda install -c pyviz bokeh cartopy datashader geoviews-core holoviews jupyter matplotlib netcdf4 numpy pandas param pyct python scipy shapely xarray

I am now working on merging install steps in the working set into maximal groups.

[jjoonathan]

Starting, as usual, from sudo docker run -it continuumio/anaconda3,

This fails at Solving Environment:

conda install -yc pyviz bokeh cartopy

This succeeds:

conda install -yc pyviz geoviews-core bokeh datashader xarray holoviews jupyter matplotlib numpy pandas param netcdf4 python scipy
conda install -yc pyviz cartopy

Huh?

[jjoonathan]

It seems I must have prematurely assumed it would fail to solve once it reported that the initial solve failed. That was incorrect. Starting from sudo docker run -it continuumio/anaconda3,

This succeeds: conda install -yc pyviz geoviews-core bokeh datashader xarray holoviews jupyter matplotlib numpy pandas param netcdf4 python scipy pyct shapely cartopy

(it is the full list of geoviews dependencies less gdal and geopandas)

It just has to go through a long process:

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): done
Solving environment: done

[jjoonathan]

gdal fails on its own. Starting from sudo docker run -it continuumio/anaconda3

we get:

(base) root@f3930e25dcba:/# conda install gdal
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): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: / 
Found conflicts! Looking for incompatible packages.
This can take several minutes.  Press CTRL-C to abort.
(... last time this took 18 hours, I'm not waiting that long to post ...)

[jjoonathan]

geopandas also fails on its own. Starting from sudo docker run -it continuumio/anaconda3

we get:

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): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: / 
Found conflicts! Looking for incompatible packages.
This can take several minutes.  Press CTRL-C to abort.
(... last time this took 18 hours, I'm not waiting that long to post ...)

[jjoonathan]

Starting from sudo docker run -it continuumio/anaconda3

This works, albeit every step is very slow:

conda update anaconda
conda install gdal
conda install geopandas

I'm now going to see if conda update anaconda fixes conda install -c pyviz geoviews.

[jjoonathan]

Indeed, it does. Starting from sudo docker run -it continuumio/anaconda3, this works:

conda update anaconda
conda install -c pyviz geoviews

It seems reasonably clear at this point that this is not a geoviews issue so much as a conda ecosystem issue. I'll close the thread.

[jjoonathan]

Actually, scratch that, should conda update anaconda be added to the README? As it stands, following the README installation instructions does not work on linux.

[philippjfr]

Sorry for all the pain you clearly had to go through here. I'd be happy to accept a PR that suggests that as a fix and lists a minimum conda version.

[jjoonathan]

Unfortunately it looks like the anaconda version that doesn't work has a name ("2020.02") while the anaconda "version" that does work does not have a name. The only concise workaround might be running an update command, which in the future may well have as much chance of breaking things as fixing them. See below.

Ok, so immediately after sudo docker run -it continuumio/anaconda3

I get:

(base) root@c1d5e120a425:/# conda list anaconda
anaconda                  2020.02                  py37_0  
anaconda-client           1.7.2                    py37_0  
anaconda-navigator        1.9.12                   py37_0  
anaconda-project          0.8.4                      py_0  

Attempting to install geoviews results in Solving Environment forever.
When I run conda update anaconda, it gives this warning:

The following packages will be DOWNGRADED:
  anaconda                                   2020.02-py37_0 --> custom-py37_1

although most (all?) of the sub packages that I can see are actually being upgraded. Full logs attached
. I guess anaconda update anaconda now means approximately what anaconda update --all used to mean, which leads to an undetermined anaconda version, which leads to poor ordering, which leads to update causing a "downgrade"? In any case, after running that command, conda install -c pyviz geoviews works and versions are listed as follows:

(base) root@c1d5e120a425:/# conda list anaconda
_anaconda_depends         2019.03                  py37_0  
anaconda                  custom                   py37_1  
anaconda-client           1.7.2                    py37_0  
anaconda-navigator        1.9.12                   py37_0  
anaconda-project          0.8.4                      py_0  

I am now testing conda update --all, the command recommended here for possible inclusion in the README.