[Bug]: ImportError: Numba needs NumPy 2.2 or less. Got NumPy 2.3.

[jypeter]

What happened?

I have at last updated my python environment on 2 different servers by removing the numpy<2 constraint. I got, at last, the latest xcdat (0.9.0) and xesmf (0.8.10) versions and asked one of our PhD students to test my environment

He came back right away telling me that he could not import xcdat, with a numba error that I can indeed reproduce (on both servers) : ImportError: Numba needs NumPy 2.2 or less. Got NumPy 2.3

Full traceback

 >python
Python 3.12.4 | packaged by conda-forge | (main, Jun 17 2024, 10:23:07) [GCC 12.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import xcdat
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/share/unix_files/cdat/miniconda3_2024-03/envs/bauer_env/lib/python3.12/site-packages/xcdat/__init__.py", line 13, in <module>
    from xcdat.regridder.accessor import RegridderAccessor  # noqa: F401
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/share/unix_files/cdat/miniconda3_2024-03/envs/bauer_env/lib/python3.12/site-packages/xcdat/regridder/__init__.py", line 1, in <module>
    from xcdat.regridder.accessor import RegridderAccessor  # noqa: F401
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/share/unix_files/cdat/miniconda3_2024-03/envs/bauer_env/lib/python3.12/site-packages/xcdat/regridder/accessor.py", line 6, in <module>
    from xcdat.regridder import regrid2, xesmf, xgcm
  File "/home/share/unix_files/cdat/miniconda3_2024-03/envs/bauer_env/lib/python3.12/site-packages/xcdat/regridder/regrid2.py", line 4, in <module>
    import sparse as sp
  File "/home/share/unix_files/cdat/miniconda3_2024-03/envs/bauer_env/lib/python3.12/site-packages/sparse/__init__.py", line 48, in <module>
    from sparse.numba_backend import *  # noqa: F403
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/share/unix_files/cdat/miniconda3_2024-03/envs/bauer_env/lib/python3.12/site-packages/sparse/numba_backend/__init__.py", line 85, in <module>
    from ._common import (
  File "/home/share/unix_files/cdat/miniconda3_2024-03/envs/bauer_env/lib/python3.12/site-packages/sparse/numba_backend/_common.py", line 8, in <module>
    import numba
  File "/home/share/unix_files/cdat/miniconda3_2024-03/envs/bauer_env/lib/python3.12/site-packages/numba/__init__.py", line 59, in <module>
    _ensure_critical_deps()
  File "/home/share/unix_files/cdat/miniconda3_2024-03/envs/bauer_env/lib/python3.12/site-packages/numba/__init__.py", line 45, in _ensure_critical_deps
    raise ImportError(msg)
ImportError: Numba needs NumPy 2.2 or less. Got NumPy 2.3.

What did you expect to happen? Are there are possible answers you came across?

Of course, I expected no error!

This seems very similar to ([Bug]: ImportError: Numba needs NumPy 2.1 or less. Got NumPy 2.2.
#744), with more recent version numbers

Are we shooting at a moving target? I will downgrade numpy with conda install "numpy<=2.2" as soon as I'm through with this bug report, but somebody (from numba?) should address this

Minimal Complete Verifiable Example (MVCE)

import xcdat

Relevant log output

Anything else we need to know?

No response

Environment

xCDAT = 0.9.0

>>> xr.show_versions()

INSTALLED VERSIONS
------------------
commit: None
python: 3.12.4 | packaged by conda-forge | (main, Jun 17 2024, 10:23:07) [GCC 12.3.0]
python-bits: 64
OS: Linux
OS-release: 3.10.0-1160.119.1.el7.x86_64
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: None
LANG: en_US.UTF-8
LOCALE: ('en_US', 'UTF-8')
libhdf5: 1.14.3
libnetcdf: 4.9.2

xarray: 2025.6.1
pandas: 2.3.0
numpy: 2.3.1
scipy: 1.15.2
netCDF4: 1.7.2
pydap: None
h5netcdf: None
h5py: None
zarr: 3.0.9
cftime: 1.6.4
nc_time_axis: None
iris: None
bottleneck: 1.5.0
dask: 2025.5.1
distributed: 2025.5.1
matplotlib: 3.10.3
cartopy: 0.24.0
seaborn: None
numbagg: None
fsspec: 2025.5.1
cupy: None
pint: 0.24.4
sparse: None
flox: None
numpy_groupies: 0.11.3
setuptools: 80.9.0
pip: 25.1.1
conda: None
pytest: None
mypy: None
IPython: 9.4.0
sphinx: None

[jypeter]

For the record, downgrading numpy had the expected result. Could somebody check what people are doing at numba that produces this recurring error?

 >conda install "numpy<=2.2"
Channels:
 - conda-forge
 - defaults
Platform: linux-64
Collecting package metadata (repodata.json): done
Solving environment: done

## Package Plan ##

  environment location: /home/share/unix_files/cdat/miniconda3_2024-03/envs/bauer_env

  added / updated specs:
    - numpy[version='<=2.2']


The following packages will be downloaded:

    package                    |            build
    ---------------------------|-----------------
    numpy-2.2.0                |  py312h7e784f5_0         8.0 MB  conda-forge
    ------------------------------------------------------------
                                           Total:         8.0 MB

The following packages will be DOWNGRADED:

  numpy                               2.3.1-py312h6cf2f7f_0 --> 2.2.0-py312h7e784f5_0


Proceed ([y]/n)? y


Downloading and Extracting Packages:

Preparing transaction: done
Verifying transaction: done
Executing transaction: done
(bauer_env) jypeter@obelix5 - ...jypeter - 70 >python
Python 3.12.4 | packaged by conda-forge | (main, Jun 17 2024, 10:23:07) [GCC 12.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import xcdat
>>>

[tomvothecoder]

Thank @jypeter for reporting this. This is a numba issue.

Can you make sure your environments are using the latest version of numba=0.61.2 with a patch to fix this issue? This patch version was released ~3 weeks ago.

If you notice in my environment below, numba=0.61.2 now ends with _1, which includes the patch to fix this issue.

conda create -c conda-forge -n xcdat_090 xcdat=0.9.0
...
The following NEW packages will be INSTALLED:
....
  numba              conda-forge/osx-arm64::numba-0.61.2-py313h2c0ffef_1 
....

More Context

This issue is related to numba conda release not properly constraining numpy <2.3 (reported here and here). There was a numba-feedstock PR to apply this constrain here, which should be available through numba=0.61.2 as a patch build 1 (e.g., conda-forge/osx-arm64::numba-0.61.2-py313h2c0ffef_1)

As you can see below, this patch version of numba=0.61.2 no longer allows numpy>2.23:

vo13@ml-5556678 ~ % conda create -c conda-forge -n numba numba numpy=2.3.0
Channels:
 - conda-forge
 - defaults
Platform: osx-arm64
Collecting package metadata (repodata.json): done
Solving environment: failed

LibMambaUnsatisfiableError: Encountered problems while solving:
  - package numba-0.56.3-py38h9197a36_0 requires numpy >=1.19,!=1.22.0,!=1.22.1,!=1.22.2,<1.23, but none of the providers can be installed

Could not solve for environment specs
The following packages are incompatible
├─ numba is installable with the potential options
....
│  └─ numba 0.61.2 would require
│     └─ numpy >=1.24,<2.3,!=2.0.1 , which can be installed;
└─ numpy 2.3.0**  is not installable because it conflicts with any installable versions previously reported.

Next Steps

If you confirm the latest version of numba=0.61.2 fixes this issue, I will close this ticket.

[jypeter]

Thanks @tomvothecoder

In the environment described above, where I downgraded numpy using conda install "numpy<=2.2", I seem to have the non-patched numba, and updating this environment without specifying a constraint would trigger reinstalling numpy 2.3.1

(bauer_env) jypeter@obelix5 - ...jypeter - 43 >conda list | grep numba
290:numba                                   0.61.2           py312h2e6246c_0                                                       conda-forge

(bauer_env) jypeter@obelix5 - ...jypeter - 45 >conda update --all
Retrieving notices: done
Channels:
 - conda-forge
 - defaults
Platform: linux-64
Collecting package metadata (repodata.json): done
Solving environment: done

## Package Plan ##

  environment location: /home/share/unix_files/cdat/miniconda3_2024-03/envs/bauer_env


The following packages will be downloaded:

    package                    |            build
    ---------------------------|-----------------
    aws-c-event-stream-0.5.5   |       h76f0014_0          57 KB  conda-forge
    aws-crt-cpp-0.32.10        |       ha543af7_2         391 KB  conda-forge
    aws-sdk-cpp-1.11.510       |      hf18ad05_13         3.2 MB  conda-forge
    binutils-2.44              |       h4852527_0          34 KB  conda-forge
    binutils_impl_linux-64-2.44|       h4bf12b8_0         3.4 MB  conda-forge
    binutils_linux-64-2.44     |       h4852527_0          35 KB  conda-forge
    fonttools-4.58.5           |  py312h178313f_0         2.7 MB  conda-forge
    ld_impl_linux-64-2.44      |       h1423503_0         660 KB  conda-forge
    libpng-1.6.50              |       h943b412_0         282 KB  conda-forge
    ruff-0.12.2                |       hcc1af86_0         8.9 MB  conda-forge
    xarray-2025.7.0            |     pyhd8ed1ab_0         858 KB  conda-forge
    zarr-3.0.10                |     pyhd8ed1ab_0         215 KB  conda-forge
    ------------------------------------------------------------
                                           Total:        20.8 MB

The following packages will be UPDATED:

  aws-c-event-stream                      0.5.4-h76f0014_12 --> 0.5.5-h76f0014_0
  aws-crt-cpp                            0.32.10-hff780f1_1 --> 0.32.10-ha543af7_2
  aws-sdk-cpp                          1.11.510-h937e755_12 --> 1.11.510-hf18ad05_13
  binutils                                  2.43-h4852527_5 --> 2.44-h4852527_0
  binutils_impl_lin~                        2.43-h4bf12b8_5 --> 2.44-h4bf12b8_0
  binutils_linux-64                         2.43-h4852527_5 --> 2.44-h4852527_0
  fonttools                          4.58.4-py312h178313f_0 --> 4.58.5-py312h178313f_0
  ld_impl_linux-64                          2.43-h712a8e2_5 --> 2.44-h1423503_0
  libpng                                  1.6.49-h943b412_0 --> 1.6.50-h943b412_0
  numpy                               2.2.0-py312h7e784f5_0 --> 2.3.1-py312h6cf2f7f_0
  ruff                                    0.12.1-hcc1af86_1 --> 0.12.2-hcc1af86_0
  xarray                              2025.6.1-pyhd8ed1ab_1 --> 2025.7.0-pyhd8ed1ab_0
  zarr                                   3.0.9-pyhd8ed1ab_0 --> 3.0.10-pyhd8ed1ab_0


Proceed ([y]/n)? n

But I have have just seen that changing my constraint from numpy<=2.2 to numpy<2.3 installs the patched numba and numpy 2.2.6

(bauer_env) jypeter@obelix5 - ...jypeter - 46 >conda update --all 'numpy<2.3'
Channels:
 - conda-forge
 - defaults
Platform: linux-64
Collecting package metadata (repodata.json): done
Solving environment: done

## Package Plan ##

  environment location: /home/share/unix_files/cdat/miniconda3_2024-03/envs/bauer_env

  added / updated specs:
    - numpy[version='<2.3']


The following packages will be downloaded:

    package                    |            build
    ---------------------------|-----------------
    aws-c-event-stream-0.5.5   |       h76f0014_0          57 KB  conda-forge
    aws-crt-cpp-0.32.10        |       ha543af7_2         391 KB  conda-forge
    aws-sdk-cpp-1.11.510       |      hf18ad05_13         3.2 MB  conda-forge
    binutils-2.44              |       h4852527_0          34 KB  conda-forge
    binutils_impl_linux-64-2.44|       h4bf12b8_0         3.4 MB  conda-forge
    binutils_linux-64-2.44     |       h4852527_0          35 KB  conda-forge
    fonttools-4.58.5           |  py312h178313f_0         2.7 MB  conda-forge
    ld_impl_linux-64-2.44      |       h1423503_0         660 KB  conda-forge
    libpng-1.6.50              |       h943b412_0         282 KB  conda-forge
    numba-0.61.2               |  py312h7bcfee6_1         5.5 MB  conda-forge
    numpy-2.2.6                |  py312h72c5963_0         8.1 MB  conda-forge
    ruff-0.12.2                |       hcc1af86_0         8.9 MB  conda-forge
    xarray-2025.7.0            |     pyhd8ed1ab_0         858 KB  conda-forge
    zarr-3.0.10                |     pyhd8ed1ab_0         215 KB  conda-forge
    ------------------------------------------------------------
                                           Total:        34.4 MB

The following packages will be UPDATED:

  aws-c-event-stream                      0.5.4-h76f0014_12 --> 0.5.5-h76f0014_0
  aws-crt-cpp                            0.32.10-hff780f1_1 --> 0.32.10-ha543af7_2
  aws-sdk-cpp                          1.11.510-h937e755_12 --> 1.11.510-hf18ad05_13
  binutils                                  2.43-h4852527_5 --> 2.44-h4852527_0
  binutils_impl_lin~                        2.43-h4bf12b8_5 --> 2.44-h4bf12b8_0
  binutils_linux-64                         2.43-h4852527_5 --> 2.44-h4852527_0
  fonttools                          4.58.4-py312h178313f_0 --> 4.58.5-py312h178313f_0
  ld_impl_linux-64                          2.43-h712a8e2_5 --> 2.44-h1423503_0
  libpng                                  1.6.49-h943b412_0 --> 1.6.50-h943b412_0
  numba                              0.61.2-py312h2e6246c_0 --> 0.61.2-py312h7bcfee6_1
  numpy                               2.2.0-py312h7e784f5_0 --> 2.2.6-py312h72c5963_0
  ruff                                    0.12.1-hcc1af86_1 --> 0.12.2-hcc1af86_0
  xarray                              2025.6.1-pyhd8ed1ab_1 --> 2025.7.0-pyhd8ed1ab_0
  zarr                                   3.0.9-pyhd8ed1ab_0 --> 3.0.10-pyhd8ed1ab_0


Proceed ([y]/n)? y

And I can still import xcdat

(bauer_env) jypeter@obelix5 - ...jypeter - 48 >conda list | grep num
numba                                   0.61.2           py312h7bcfee6_1          conda-forge
numpy                                   2.2.6            py312h72c5963_0          conda-forge
[... and some removed *num* stuff ... ]
(bauer_env) jypeter@obelix5 - ...jypeter - 49 >
(bauer_env) jypeter@obelix5 - ...jypeter - 49 >python
Python 3.12.4 | packaged by conda-forge | (main, Jun 17 2024, 10:23:07) [GCC 12.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import xcdat
>>>

But then an horror story starts! If I try to update this environment, expecting conda to tell me that there is nothing to do, because I have just updated my environment a few minutes ago, it now

• wants to update numpy to 2.3.1
• wants to remove numba and xesmf
• wants to seriously downgrade xcdat and xskillscore

Of course I don't want to proceed!

(bauer_env) jypeter@obelix5 - ...jypeter - 50 >conda update --all
Channels:
 - conda-forge
 - defaults
Platform: linux-64
Collecting package metadata (repodata.json): done
Solving environment: done

## Package Plan ##

  environment location: /home/share/unix_files/cdat/miniconda3_2024-03/envs/bauer_env


The following packages will be downloaded:

    package                    |            build
    ---------------------------|-----------------
    lxml-6.0.0                 |  py312h68d7fa5_0         1.5 MB  conda-forge
    xcdat-0.6.0                |     pyhd8ed1ab_0          71 KB  conda-forge
    xskillscore-0.0.5          |             py_0          16 KB  conda-forge
    ------------------------------------------------------------
                                           Total:         1.6 MB

The following NEW packages will be INSTALLED:

  libxslt            conda-forge/linux-64::libxslt-1.1.39-h76b75d6_0
  lxml               conda-forge/linux-64::lxml-6.0.0-py312h68d7fa5_0

The following packages will be REMOVED:

  bottleneck-1.5.0-py312hc0a28a1_0
  esmf-8.8.1-nompi_h4441c20_0
  esmpy-8.8.1-pyhecae5ae_0
  llvmlite-0.44.0-py312h374181b_1
  netcdf-fortran-4.6.1-nompi_h22f9119_108
  numba-0.61.2-py312h7bcfee6_1
  sparse-0.17.0-pyhcf101f3_0
  xesmf-0.8.10-pyhd8ed1ab_0
  xhistogram-0.3.2-pyhd8ed1ab_0

The following packages will be UPDATED:

  numpy                               2.2.6-py312h72c5963_0 --> 2.3.1-py312h6cf2f7f_0

The following packages will be DOWNGRADED:

  xcdat                                  0.9.0-pyhd8ed1ab_0 --> 0.6.0-pyhd8ed1ab_0
  xskillscore                           0.0.26-pyhd8ed1ab_1 --> 0.0.5-py_0


Proceed ([y]/n)? n

On the other hand, if I still constrain numpy during the update, it tells me I'm OK

(bauer_env) jypeter@obelix5 - ...jypeter - 51 >conda update --all 'numpy<2.3'
Channels:
 - conda-forge
 - defaults
Platform: linux-64
Collecting package metadata (repodata.json): done
Solving environment: done

# All requested packages already installed.

Is the patched numba really fixing something?

[tomvothecoder]

Hey @xylar, sorry to bug you but I wanted to check if I'm on the right track here.

In my comment, I noted that the issue appears to stem from build 0 of numba=0.61.2 not enforcing a numpy <2.3 constraint, which led to compatibility problems. A patch was introduced in build 1 of numba=0.61.2, which properly prevents it from being installed with numpy >=2.3.

Now, @jypeter is running into trouble when using conda update --all, which attempts to:

• upgrade numpy from 2.2.6 to 2.3.1,
• remove numba, xesmf, etc.,
• and downgrade xcdat, xskillscore, and other packages.

This seems to be the conda solver aggressively trying to update all packages, ignoring the compatibility boundaries unless they're explicitly pinned. When he runs conda update --all 'numpy<2.3', everything resolves cleanly.

Would you say it's more advisable to only update specific major packages individually (e.g., xarray, xcdat, etc.) rather than using conda update --all? Or even re-creating the environment from scratch with up-to-date but compatible versions?

[xylar]

In my #775 (comment), I noted that the issue appears to stem from build 0 of numba=0.61.2 not enforcing a numpy <2.3 constraint, which led to compatibility problems. A patch was introduced in build 1 of numba=0.61.2, which properly prevents it from being installed with numpy >=2.3.

Someone probably needs to make a PR to https://github.com/conda-forge/admin-requests to fix the version constrain in build 0. Otherwise, that build will still remain a valid alternative. build 1 gets favored over build 0 but that doesn't mean the solver won't fall back on build 0 if it helps solve the whole matrix of dependencies.

Would you say it's more advisable to only update specific major packages individually (e.g., xarray, xcdat, etc.) rather than using conda update --all? Or even re-creating the environment from scratch with up-to-date but compatible versions?

I personally onlu use conda update --all on my base environment, which only has some basics like conda build python, pip and such.

I would always recommend using conda create or conda env create to make a fresh environment over trying to update packages. I don't think that will necessarily save you from this particular issue. But you will nearly always get a much cleaner, more optimal solve with conda create than conda update.

[tomvothecoder]

Someone probably needs to make a PR to conda-forge/admin-requests to fix the version constrain in build 0. Otherwise, that build will still remain a valid alternative. build 1 gets favored over build 0 but that doesn't mean the solver won't fall back on build 0 if it helps solve the whole matrix of dependencies.

Thank you @xylar. I opened a GitHub issue directly on the conda-forge/numba-feedstock here. Hopefully a feedstock maintainer can handle this for us.

@jypeter, for now, I would follow Xylar's advice in the comment above to create fresh environment with conda create/conda env create. Unfortunately, this seems like a persistent issue that gets missed when numba is released on conda-forge and it is out of our control. We can explicitly pin numpy <2.3 in xCDAT, but I prefer not doing that if numba feedstock maintainers can address it directly.

[xylar]

As I commented in conda-forge/numba-feedstock#159 (comment), I think this has already been addressed in conda-forge/conda-forge-repodata-patches-feedstock#1039 (I linked to the wrong repo above for fixing the issue).

I don't think there's anything further than can be done on conda-forge about this and it's just a question of making sure you get the latest, patched version of the package from conda-forge. This might require cleaning out old cached packages. Or a fresh install with conda create should take care of it.

[tomvothecoder]

As I commented in conda-forge/numba-feedstock#159 (comment), I think this has already been addressed in conda-forge/conda-forge-repodata-patches-feedstock#1039 (I linked to the wrong repo above for fixing the issue).

I don't think there's anything further than can be done on conda-forge about this and it's just a question of making sure you get the latest, patched version of the package from conda-forge. This might require cleaning out old cached packages. Or a fresh install with conda create should take care of it.

@jypeter can you try cleaning out your cached packages and/or re-create the environment from scratch? Hopefully this solves your issue.

[jypeter]

Thanks @tomvothecoder and @xylar for investigating this. I will check tomorrow (hopefully) what happens if I clear the cache before updating the existing environment

I'd rather rely on package dependencies specified correctly, and conda (now that it relies on the mamba solver by default), rather than having to regularly create an environment from scratch (even if I should also experiment with this)

I'm trying to maintain one environment used my multiple users, who need different sets of packages, and it fortunately works fine most of the time (thanks to the mamba-based solver, and using only packages from conda-forge)

In the worst case scenario, I can probably direct the users to a brand new environment without them having to change anything, but I will try other options first

[tomvothecoder]

Thanks @tomvothecoder and @xylar for investigating this. I will check tomorrow (hopefully) what happens if I clear the cache before updating the existing environment

@jypeter Any luck?

I'd rather rely on package dependencies specified correctly, and conda (now that it relies on the mamba solver by default), rather than having to regularly create an environment from scratch (even if I should also experiment with this)

I'm trying to maintain one environment used my multiple users, who need different sets of packages, and it fortunately works fine most of the time (thanks to the mamba-based solver, and using only packages from conda-forge)

Got it, thanks for the clarification on your environment setup and management process. @xylar has been in charge of E3SM Unified, which a metapackage for a unified conda environment for analysis and other pre- and post-processing for the Energy Exascale Earth System Model (E3SM).

E3SM Unified environments are versioned, which helps create reproducible and maintainable environments. Maybe following a similar approach would be helpful (for your case, as updating all packages within a single, shared environment can easily introduce side-effects, bugs, etc. (e.g., like this numba error).

In the worst case scenario, I can probably direct the users to a brand new environment without them having to change anything, but I will try other options first

Alternatively, you can re-create the environment from scratch with the same name and the same shared directory. This would allow users to re-use the same environment with all the updates.

[jypeter]

Sorry for the delay! I have just started experimenting again, on two different clusters, where I maintain very similar environments. When I need to mention the clusters, I will use:

• cluster O (our local obelix cluster, using my bauer_env environment). Everything I have written above was done on cluster O, but I had the same behavior on the other cluster

  • numba 0.61.2 py312h7bcfee6_1 conda-forge
numpy 2.2.6 py312h72c5963_0 conda-forge
xcdat 0.9.0 pyhd8ed1ab_0 conda-forge

• cluster S (our remotte spirit cluster, using my bezaud_env environment). The numpy version is slightly older

  • numba 0.61.2 py312h7bcfee6_1 conda-forge
numpy 2.2.0 py312h7e784f5_0 conda-forge
xcdat 0.9.0 pyhd8ed1ab_0 conda-forge

I can import xcdat correctly in both environments

Two days ago, I updated the base environment on cluster O, and as far as I remember, conda update -n bauer_env --all still wanted to remove numba and downgrade xcdat the way it is described above, and I aborted the update

Earlier today, I updated the base environment on both clusters, and I have conda 25.5.1 on both clusters

Today's base update output on cluster O:

>conda update --all
[...]
The following packages will be downloaded:

    package                    |            build
    ---------------------------|-----------------
    expat-2.7.1                |       hecca717_0         138 KB  conda-forge
    libexpat-2.7.1             |       hecca717_0          73 KB  conda-forge
    libsqlite-3.50.3           |       hee844dc_0         914 KB  conda-forge
    python_abi-3.12            |          8_cp312           7 KB  conda-forge
    sqlite-3.50.3              |       heff268d_0         162 KB  conda-forge
    ------------------------------------------------------------
                                           Total:         1.3 MB

The following packages will be UPDATED:

  expat                                    2.7.0-h5888daf_0 --> 2.7.1-hecca717_0
  libexpat                                 2.7.0-h5888daf_0 --> 2.7.1-hecca717_0
  libsqlite                               3.50.2-hee844dc_2 --> 3.50.3-hee844dc_0
  python_abi                                   3.12-7_cp312 --> 3.12-8_cp312
  sqlite                                  3.50.2-heff268d_2 --> 3.50.3-heff268d_0


Proceed ([y]/n)? y

And now, nothing suspicious happens if I update cluster O: numpy and numba are left unchanged, and I get lots of updates (including xcdat 0.9.0 => 0.9.1)

 >conda update -n bauer_env --all
[...]
The following packages will be downloaded:

    package                    |            build
    ---------------------------|-----------------
    aws-c-auth-0.9.0           |      hcc895bc_17         120 KB  conda-forge
    aws-c-cal-0.9.2            |       he7b75e1_1          50 KB  conda-forge
    aws-c-common-0.12.4        |       hb03c661_0         231 KB  conda-forge
    aws-c-compression-0.3.1    |       h92c474e_6          22 KB  conda-forge
    aws-c-event-stream-0.5.5   |       h84d2157_2          57 KB  conda-forge
    aws-c-http-0.10.3          |       hbe0f4a8_1         219 KB  conda-forge
    aws-c-io-0.21.1            |       h1d8da38_0         175 KB  conda-forge
    aws-c-mqtt-0.13.3          |       h9c20097_1         211 KB  conda-forge
    aws-c-s3-0.8.5             |       h317ce67_1         132 KB  conda-forge
    aws-c-sdkutils-0.2.4       |       h92c474e_1          58 KB  conda-forge
    aws-checksums-0.2.7        |       h92c474e_2          75 KB  conda-forge
    aws-crt-cpp-0.33.0         |       h4f272d1_2         397 KB  conda-forge
    aws-sdk-cpp-1.11.510       |      h455e09b_15         3.3 MB  conda-forge
    azure-core-cpp-1.15.0      |       h5cfcd09_0         338 KB  conda-forge
    azure-identity-cpp-1.11.0  |       hb5324b0_1         228 KB  conda-forge
    azure-storage-blobs-cpp-12.13.0|       hf182047_2         538 KB  conda-forge
    azure-storage-common-cpp-12.10.0|       h40e822a_1         147 KB  conda-forge
    azure-storage-files-datalake-cpp-12.12.0|       h141ff2a_2         280 KB  conda-forge
    binutils-2.44              |       h4852527_1          34 KB  conda-forge
    binutils_impl_linux-64-2.44|       h4bf12b8_1         3.6 MB  conda-forge
    binutils_linux-64-2.44     |       h4852527_1          35 KB  conda-forge
    dask-2025.7.0              |     pyhe01879c_0          11 KB  conda-forge
    dask-core-2025.7.0         |     pyhe01879c_1         1.0 MB  conda-forge
    debugpy-1.8.15             |  py312h8285ef7_0         2.7 MB  conda-forge
    distributed-2025.7.0       |     pyhe01879c_0         828 KB  conda-forge
    fonttools-4.59.0           |  py312h8a5da7c_0         2.7 MB  conda-forge
    fsspec-2025.7.0            |     pyhd8ed1ab_0         142 KB  conda-forge
    jsonschema-4.24.1          |     pyhe01879c_0          80 KB  conda-forge
    jsonschema-with-format-nongpl-4.24.1|       he01879c_0           5 KB  conda-forge
    libabseil-20250512.1       | cxx17_hba17884_0         1.2 MB  conda-forge
    libarrow-20.0.0            |  h94aa55a_14_cpu         9.0 MB  conda-forge
    libarrow-acero-20.0.0      |  h635bf11_14_cpu         644 KB  conda-forge
    libarrow-dataset-20.0.0    |  h635bf11_14_cpu         613 KB  conda-forge
    libarrow-substrait-20.0.0  |  h3f74fd7_14_cpu         504 KB  conda-forge
    libgoogle-cloud-2.39.0     |       hdb79228_0         1.2 MB  conda-forge
    libgoogle-cloud-storage-2.39.0|       hdbdcf42_0         785 KB  conda-forge
    libgrpc-1.73.1             |       h1e535eb_0         8.0 MB  conda-forge
    libopentelemetry-cpp-1.21.0|       hb9b0907_1         865 KB  conda-forge
    libopentelemetry-cpp-headers-1.21.0|       ha770c72_1         355 KB  conda-forge
    libparquet-20.0.0          |  h790f06f_14_cpu         1.2 MB  conda-forge
    libprotobuf-6.31.1         |       h9ef548d_1         3.8 MB  conda-forge
    libre2-11-2025.06.26       |       h7064273_1         205 KB  conda-forge
    libthrift-0.22.0           |       h093b73b_0         416 KB  conda-forge
    libwebp-base-1.6.0         |       hd42ef1d_0         419 KB  conda-forge
    multiurl-0.3.6             |     pyhd8ed1ab_0          22 KB  conda-forge
    narwhals-1.47.1            |     pyhe01879c_0         232 KB  conda-forge
    orc-2.1.3                  |       h61e0c1e_0         1.2 MB  conda-forge
    pandas-2.3.1               |  py312hf79963d_0        14.4 MB  conda-forge
    pybind11-3.0.0             |     pyh9380348_1         226 KB  conda-forge
    pybind11-global-3.0.0      |     pyhf748d72_1         222 KB  conda-forge
    re2-2025.06.26             |       h9925aae_1          27 KB  conda-forge
    ruff-0.12.4                |       hf9daec2_0         8.7 MB  conda-forge
    s2n-1.5.22                 |       h96f233e_0         349 KB  conda-forge
    scikit-learn-1.7.1         |  py312h4f0b9e3_0         9.2 MB  conda-forge
    scipy-1.16.0               |  py312hf734454_0        16.1 MB  conda-forge
    statsmodels-0.14.5         |  py312h8b63200_0        11.5 MB  conda-forge
    types-python-dateutil-2.9.0.20250708|     pyhd8ed1ab_0          24 KB  conda-forge
    typing-extensions-4.14.1   |       h4440ef1_0          88 KB  conda-forge
    typing_extensions-4.14.1   |     pyhe01879c_0          50 KB  conda-forge
    xarray-2025.7.1            |     pyhd8ed1ab_0         862 KB  conda-forge
    xcdat-0.9.1                |     pyhd8ed1ab_0          73 KB  conda-forge
    xskillscore-0.0.27         |     pyhd8ed1ab_0          68 KB  conda-forge
    zarr-3.1.0                 |     pyhe01879c_0         260 KB  conda-forge
    ------------------------------------------------------------
                                           Total:       110.5 MB

The following packages will be REMOVED:

  importlib_resources-6.5.2-pyhd8ed1ab_0
  pkgutil-resolve-name-1.3.10-pyhd8ed1ab_2

The following packages will be UPDATED:

  aws-c-auth                              0.9.0-hbfa7f16_15 --> 0.9.0-hcc895bc_17
  aws-c-cal                                0.9.2-h5e3027f_0 --> 0.9.2-he7b75e1_1
  aws-c-common                            0.12.3-hb9d3cd8_0 --> 0.12.4-hb03c661_0
  aws-c-compression                        0.3.1-hafb2847_5 --> 0.3.1-h92c474e_6
  aws-c-event-stream                       0.5.5-h76f0014_0 --> 0.5.5-h84d2157_2
  aws-c-http                              0.10.2-h015de20_2 --> 0.10.3-hbe0f4a8_1
  aws-c-io                                0.20.1-hdfce8c9_0 --> 0.21.1-h1d8da38_0
  aws-c-mqtt                              0.13.1-h1e5e6c0_3 --> 0.13.3-h9c20097_1
  aws-c-s3                                 0.8.3-h5e174a9_0 --> 0.8.5-h317ce67_1
  aws-c-sdkutils                           0.2.4-hafb2847_0 --> 0.2.4-h92c474e_1
  aws-checksums                            0.2.7-hafb2847_1 --> 0.2.7-h92c474e_2
  aws-crt-cpp                            0.32.10-ha543af7_2 --> 0.33.0-h4f272d1_2
  aws-sdk-cpp                          1.11.510-hf18ad05_13 --> 1.11.510-h455e09b_15
  azure-core-cpp                          1.14.0-h5cfcd09_0 --> 1.15.0-h5cfcd09_0
  azure-identity-cpp                      1.10.0-h113e628_0 --> 1.11.0-hb5324b0_1
  azure-storage-blo~                     12.13.0-h3cf044e_1 --> 12.13.0-hf182047_2
  azure-storage-com~                      12.8.0-h736e048_1 --> 12.10.0-h40e822a_1
  azure-storage-fil~                     12.12.0-ha633028_1 --> 12.12.0-h141ff2a_2
  binutils                                  2.44-h4852527_0 --> 2.44-h4852527_1
  binutils_impl_lin~                        2.44-h4bf12b8_0 --> 2.44-h4bf12b8_1
  binutils_linux-64                         2.44-h4852527_0 --> 2.44-h4852527_1
  ca-certificates                      2025.6.15-hbd8a1cb_0 --> 2025.7.14-hbd8a1cb_0
  certifi                            2025.6.15-pyhd8ed1ab_0 --> 2025.7.14-pyhd8ed1ab_0
  dask                                2025.5.1-pyhd8ed1ab_0 --> 2025.7.0-pyhe01879c_0
  dask-core                           2025.5.1-pyhd8ed1ab_0 --> 2025.7.0-pyhe01879c_1
  debugpy                            1.8.14-py312h2ec8cdc_0 --> 1.8.15-py312h8285ef7_0
  distributed                         2025.5.1-pyhd8ed1ab_0 --> 2025.7.0-pyhe01879c_0
  expat                                    2.7.0-h5888daf_0 --> 2.7.1-hecca717_0
  fonttools                          4.58.5-py312h178313f_0 --> 4.59.0-py312h8a5da7c_0
  fsspec                              2025.5.1-pyhd8ed1ab_0 --> 2025.7.0-pyhd8ed1ab_0
  jsonschema                            4.24.0-pyhd8ed1ab_0 --> 4.24.1-pyhe01879c_0
  jsonschema-with-f~                      4.24.0-hd8ed1ab_0 --> 4.24.1-he01879c_0
  ld_impl_linux-64                          2.44-h1423503_0 --> 2.44-h1423503_1
  libabseil                     20250127.1-cxx17_hbbce691_0 --> 20250512.1-cxx17_hba17884_0
  libarrow                            20.0.0-h1b9301b_8_cpu --> 20.0.0-h94aa55a_14_cpu
  libarrow-acero                      20.0.0-hcb10f89_8_cpu --> 20.0.0-h635bf11_14_cpu
  libarrow-dataset                    20.0.0-hcb10f89_8_cpu --> 20.0.0-h635bf11_14_cpu
  libarrow-substrait                  20.0.0-h1bed206_8_cpu --> 20.0.0-h3f74fd7_14_cpu
  libexpat                                 2.7.0-h5888daf_0 --> 2.7.1-hecca717_0
  libgoogle-cloud                         2.36.0-hc4361e1_1 --> 2.39.0-hdb79228_0
  libgoogle-cloud-s~                      2.36.0-h0121fbd_1 --> 2.39.0-hdbdcf42_0
  libgrpc                                 1.71.0-h8e591d7_1 --> 1.73.1-h1e535eb_0
  libopentelemetry-~                      1.21.0-hd1b1c89_0 --> 1.21.0-hb9b0907_1
  libopentelemetry-~                      1.21.0-ha770c72_0 --> 1.21.0-ha770c72_1
  libparquet                          20.0.0-h081d1f1_8_cpu --> 20.0.0-h790f06f_14_cpu
  libprotobuf                             5.29.3-h501fc15_1 --> 6.31.1-h9ef548d_1
  libre2-11                           2025.06.26-hba17884_0 --> 2025.06.26-h7064273_1
  libsqlite                               3.50.2-h6cd9bfd_0 --> 3.50.3-hee844dc_0
  libthrift                               0.21.0-h0e7cc3e_0 --> 0.22.0-h093b73b_0
  libwebp-base                             1.5.0-h851e524_0 --> 1.6.0-hd42ef1d_0
  multiurl                               0.3.5-pyhd8ed1ab_0 --> 0.3.6-pyhd8ed1ab_0
  narwhals                              1.45.0-pyhe01879c_0 --> 1.47.1-pyhe01879c_0
  orc                                      2.1.2-h17f744e_0 --> 2.1.3-h61e0c1e_0
  pandas                              2.3.0-py312hf9745cd_0 --> 2.3.1-py312hf79963d_0
  pybind11                              2.13.6-pyhc790b64_3 --> 3.0.0-pyh9380348_1
  pybind11-global                       2.13.6-pyh217bc35_3 --> 3.0.0-pyhf748d72_1
  python_abi                                   3.12-7_cp312 --> 3.12-8_cp312
  re2                                 2025.06.26-h9925aae_0 --> 2025.06.26-h9925aae_1
  ruff                                    0.12.2-hcc1af86_0 --> 0.12.4-hf9daec2_0
  s2n                                     1.5.21-h7ab7c64_0 --> 1.5.22-h96f233e_0
  scikit-learn                        1.7.0-py312h7a48858_1 --> 1.7.1-py312h4f0b9e3_0
  scipy                              1.15.2-py312ha707e6e_0 --> 1.16.0-py312hf734454_0
  sqlite                                  3.50.2-hb7a22d2_0 --> 3.50.3-heff268d_0
  statsmodels                        0.14.4-py312h8b63200_2 --> 0.14.5-py312h8b63200_0
  types-python-date~            2.9.0.20250516-pyhd8ed1ab_0 --> 2.9.0.20250708-pyhd8ed1ab_0
  typing-extensions                       4.14.0-h32cad80_0 --> 4.14.1-h4440ef1_0
  typing_extensions                     4.14.0-pyhe01879c_0 --> 4.14.1-pyhe01879c_0
  xarray                              2025.7.0-pyhd8ed1ab_0 --> 2025.7.1-pyhd8ed1ab_0
  xcdat                                  0.9.0-pyhd8ed1ab_0 --> 0.9.1-pyhd8ed1ab_0
  xskillscore                           0.0.26-pyhd8ed1ab_1 --> 0.0.27-pyhd8ed1ab_0
  zarr                                  3.0.10-pyhd8ed1ab_0 --> 3.1.0-pyhe01879c_0

Proceed ([y]/n)? y

Fortunately, nothing obvious seems broken. At least I can still import (the new) xcdat !

I don't know why the update worked this time. Note that I did not have to use conda clean --all on cluster O in order to make things work

On the other hand, I was out of (files) quota on cluster S and I had to use conda clean --all in order to just update conda! conda can indeed get hungry

[...]
   OSError: [Errno 122] Disk quota exceeded: '/homedata/jypmce/miniconda3_2024-03/pkgs/cache/497deca9.1c00.tmp'
[...]
$ quotas
Quota for user jypmce@spiritx
     Filesystem    used   quota   limit   grace   files   quota   limit   grace
[...]
      /homedata  14.26G      1T    1.1T       -  327878* 300000  350000    none
[...]
(base) jypmce@spiritx1:~$ conda clean --all
Will remove 970 (1.74 GB) tarball(s).
Proceed ([y]/n)? y

Will remove 1 index cache(s).
Proceed ([y]/n)? y

Will remove 856 (6.51 GB) package(s).
Proceed ([y]/n)? y

There are no tempfile(s) to remove.
There are no logfile(s) to remove.

$ quotas
Quota for user jypmce@spiritx
     Filesystem    used   quota   limit   grace   files   quota   limit   grace
      /homedata  3.922G      1T    1.1T       -  122540  300000  350000       -

The update also goes on smoothly on cluster S, with numba unchanged, and numpy and xcdat getting upgraded to the same versions as on cluster O: numpy 2.2.0 => 2.2.6 and xcdat 0.9.0 => 0.9.1. And with 20k extra files if I check the quota...

$ conda update -n bezaud_env --all
[...]
The following packages will be downloaded:

    package                    |            build
    ---------------------------|-----------------
    aws-c-auth-0.9.0           |      hcc895bc_17         120 KB  conda-forge
    aws-c-cal-0.9.2            |       he7b75e1_1          50 KB  conda-forge
    aws-c-common-0.12.4        |       hb03c661_0         231 KB  conda-forge
    aws-c-compression-0.3.1    |       h92c474e_6          22 KB  conda-forge
    aws-c-event-stream-0.5.5   |       h84d2157_2          57 KB  conda-forge
    aws-c-http-0.10.3          |       hbe0f4a8_1         219 KB  conda-forge
    aws-c-io-0.21.1            |       h1d8da38_0         175 KB  conda-forge
    aws-c-mqtt-0.13.3          |       h9c20097_1         211 KB  conda-forge
    aws-c-s3-0.8.5             |       h317ce67_1         132 KB  conda-forge
    aws-c-sdkutils-0.2.4       |       h92c474e_1          58 KB  conda-forge
    aws-checksums-0.2.7        |       h92c474e_2          75 KB  conda-forge
    aws-crt-cpp-0.33.0         |       h4f272d1_2         397 KB  conda-forge
    aws-sdk-cpp-1.11.510       |      h455e09b_15         3.3 MB  conda-forge
    azure-core-cpp-1.15.0      |       h5cfcd09_0         338 KB  conda-forge
    azure-identity-cpp-1.11.0  |       hb5324b0_1         228 KB  conda-forge
    azure-storage-blobs-cpp-12.13.0|       hf182047_2         538 KB  conda-forge
    azure-storage-common-cpp-12.10.0|       h40e822a_1         147 KB  conda-forge
    azure-storage-files-datalake-cpp-12.12.0|       h141ff2a_2         280 KB  conda-forge
    dask-2025.7.0              |     pyhe01879c_0          11 KB  conda-forge
    dask-core-2025.7.0         |     pyhe01879c_1         1.0 MB  conda-forge
    debugpy-1.8.15             |  py312h8285ef7_0         2.7 MB  conda-forge
    distributed-2025.7.0       |     pyhe01879c_0         828 KB  conda-forge
    fonttools-4.59.0           |  py312h8a5da7c_0         2.7 MB  conda-forge
    fsspec-2025.7.0            |     pyhd8ed1ab_0         142 KB  conda-forge
    jsonschema-4.24.1          |     pyhe01879c_0          80 KB  conda-forge
    jsonschema-with-format-nongpl-4.24.1|       he01879c_0           5 KB  conda-forge
    ld_impl_linux-64-2.44      |       h1423503_1         660 KB  conda-forge
    libabseil-20250512.1       | cxx17_hba17884_0         1.2 MB  conda-forge
    libarrow-20.0.0            |  h94aa55a_14_cpu         9.0 MB  conda-forge
    libarrow-acero-20.0.0      |  h635bf11_14_cpu         644 KB  conda-forge
    libarrow-dataset-20.0.0    |  h635bf11_14_cpu         613 KB  conda-forge
    libarrow-substrait-20.0.0  |  h3f74fd7_14_cpu         504 KB  conda-forge
    libgoogle-cloud-2.39.0     |       hdb79228_0         1.2 MB  conda-forge
    libgoogle-cloud-storage-2.39.0|       hdbdcf42_0         785 KB  conda-forge
    libgrpc-1.73.1             |       h1e535eb_0         8.0 MB  conda-forge
    libopentelemetry-cpp-1.21.0|       hb9b0907_1         865 KB  conda-forge
    libopentelemetry-cpp-headers-1.21.0|       ha770c72_1         355 KB  conda-forge
    libparquet-20.0.0          |  h790f06f_14_cpu         1.2 MB  conda-forge
    libprotobuf-6.31.1         |       h9ef548d_1         3.8 MB  conda-forge
    libre2-11-2025.06.26       |       h7064273_1         205 KB  conda-forge
    libthrift-0.22.0           |       h093b73b_0         416 KB  conda-forge
    libwebp-base-1.6.0         |       hd42ef1d_0         419 KB  conda-forge
    mapclassify-2.10.0         |     pyhd8ed1ab_1         792 KB  conda-forge
    narwhals-1.47.1            |     pyhe01879c_0         232 KB  conda-forge
    numpy-2.2.6                |  py312h72c5963_0         8.1 MB  conda-forge
    orc-2.1.3                  |       h61e0c1e_0         1.2 MB  conda-forge
    pandas-2.3.1               |  py312hf79963d_0        14.4 MB  conda-forge
    python_abi-3.12            |          8_cp312           7 KB  conda-forge
    re2-2025.06.26             |       h9925aae_1          27 KB  conda-forge
    s2n-1.5.22                 |       h96f233e_0         349 KB  conda-forge
    scikit-learn-1.7.1         |  py312h4f0b9e3_0         9.2 MB  conda-forge
    scipy-1.16.0               |  py312hf734454_0        16.1 MB  conda-forge
    statsmodels-0.14.5         |  py312h8b63200_0        11.5 MB  conda-forge
    types-python-dateutil-2.9.0.20250708|     pyhd8ed1ab_0          24 KB  conda-forge
    typing-extensions-4.14.1   |       h4440ef1_0          88 KB  conda-forge
    typing_extensions-4.14.1   |     pyhe01879c_0          50 KB  conda-forge
    xarray-2025.7.1            |     pyhd8ed1ab_0         862 KB  conda-forge
    xcdat-0.9.1                |     pyhd8ed1ab_0          73 KB  conda-forge
    xskillscore-0.0.27         |     pyhd8ed1ab_0          68 KB  conda-forge
    ------------------------------------------------------------
                                           Total:       106.9 MB

The following packages will be REMOVED:

  importlib_resources-6.5.2-pyhd8ed1ab_0
  pkgutil-resolve-name-1.3.10-pyhd8ed1ab_2

The following packages will be UPDATED:

  aws-c-auth                              0.9.0-hbfa7f16_15 --> 0.9.0-hcc895bc_17
  aws-c-cal                                0.9.2-h5e3027f_0 --> 0.9.2-he7b75e1_1
  aws-c-common                            0.12.3-hb9d3cd8_0 --> 0.12.4-hb03c661_0
  aws-c-compression                        0.3.1-hafb2847_5 --> 0.3.1-h92c474e_6
  aws-c-event-stream                       0.5.5-h76f0014_0 --> 0.5.5-h84d2157_2
  aws-c-http                              0.10.2-h015de20_2 --> 0.10.3-hbe0f4a8_1
  aws-c-io                                0.20.1-hdfce8c9_0 --> 0.21.1-h1d8da38_0
  aws-c-mqtt                              0.13.1-h1e5e6c0_3 --> 0.13.3-h9c20097_1
  aws-c-s3                                 0.8.3-h5e174a9_0 --> 0.8.5-h317ce67_1
  aws-c-sdkutils                           0.2.4-hafb2847_0 --> 0.2.4-h92c474e_1
  aws-checksums                            0.2.7-hafb2847_1 --> 0.2.7-h92c474e_2
  aws-crt-cpp                            0.32.10-ha543af7_2 --> 0.33.0-h4f272d1_2
  aws-sdk-cpp                          1.11.510-hf18ad05_13 --> 1.11.510-h455e09b_15
  azure-core-cpp                          1.14.0-h5cfcd09_0 --> 1.15.0-h5cfcd09_0
  azure-identity-cpp                      1.10.0-h113e628_0 --> 1.11.0-hb5324b0_1
  azure-storage-blo~                     12.13.0-h3cf044e_1 --> 12.13.0-hf182047_2
  azure-storage-com~                      12.8.0-h736e048_1 --> 12.10.0-h40e822a_1
  azure-storage-fil~                     12.12.0-ha633028_1 --> 12.12.0-h141ff2a_2
  ca-certificates                      2025.6.15-hbd8a1cb_0 --> 2025.7.14-hbd8a1cb_0
  certifi                            2025.6.15-pyhd8ed1ab_0 --> 2025.7.14-pyhd8ed1ab_0
  dask                                2025.5.1-pyhd8ed1ab_0 --> 2025.7.0-pyhe01879c_0
  dask-core                           2025.5.1-pyhd8ed1ab_0 --> 2025.7.0-pyhe01879c_1
  debugpy                            1.8.14-py312h2ec8cdc_0 --> 1.8.15-py312h8285ef7_0
  distributed                         2025.5.1-pyhd8ed1ab_0 --> 2025.7.0-pyhe01879c_0
  expat                                    2.7.0-h5888daf_0 --> 2.7.1-hecca717_0
  fonttools                          4.58.5-py312h178313f_0 --> 4.59.0-py312h8a5da7c_0
  fsspec                              2025.5.1-pyhd8ed1ab_0 --> 2025.7.0-pyhd8ed1ab_0
  jsonschema                            4.24.0-pyhd8ed1ab_0 --> 4.24.1-pyhe01879c_0
  jsonschema-with-f~                      4.24.0-hd8ed1ab_0 --> 4.24.1-he01879c_0
  ld_impl_linux-64                          2.44-h1423503_0 --> 2.44-h1423503_1
  libabseil                     20250127.1-cxx17_hbbce691_0 --> 20250512.1-cxx17_hba17884_0
  libarrow                            20.0.0-h1b9301b_8_cpu --> 20.0.0-h94aa55a_14_cpu
  libarrow-acero                      20.0.0-hcb10f89_8_cpu --> 20.0.0-h635bf11_14_cpu
  libarrow-dataset                    20.0.0-hcb10f89_8_cpu --> 20.0.0-h635bf11_14_cpu
  libarrow-substrait                  20.0.0-h1bed206_8_cpu --> 20.0.0-h3f74fd7_14_cpu
  libexpat                                 2.7.0-h5888daf_0 --> 2.7.1-hecca717_0
  libgoogle-cloud                         2.36.0-hc4361e1_1 --> 2.39.0-hdb79228_0
  libgoogle-cloud-s~                      2.36.0-h0121fbd_1 --> 2.39.0-hdbdcf42_0
  libgrpc                                 1.71.0-h8e591d7_1 --> 1.73.1-h1e535eb_0
  libopentelemetry-~                      1.21.0-hd1b1c89_0 --> 1.21.0-hb9b0907_1
  libopentelemetry-~                      1.21.0-ha770c72_0 --> 1.21.0-ha770c72_1
  libparquet                          20.0.0-h081d1f1_8_cpu --> 20.0.0-h790f06f_14_cpu
  libprotobuf                             5.29.3-h501fc15_1 --> 6.31.1-h9ef548d_1
  libre2-11                           2025.06.26-hba17884_0 --> 2025.06.26-h7064273_1
  libsqlite                               3.50.2-h6cd9bfd_0 --> 3.50.3-hee844dc_0
  libthrift                               0.21.0-h0e7cc3e_0 --> 0.22.0-h093b73b_0
  libwebp-base                             1.5.0-h851e524_0 --> 1.6.0-hd42ef1d_0
  mapclassify                            2.9.0-pyhd8ed1ab_0 --> 2.10.0-pyhd8ed1ab_1
  narwhals                              1.45.0-pyhe01879c_0 --> 1.47.1-pyhe01879c_0
  numpy                               2.2.0-py312h7e784f5_0 --> 2.2.6-py312h72c5963_0
  orc                                      2.1.2-h17f744e_0 --> 2.1.3-h61e0c1e_0
  pandas                              2.3.0-py312hf9745cd_0 --> 2.3.1-py312hf79963d_0
  python_abi                                   3.12-7_cp312 --> 3.12-8_cp312
  re2                                 2025.06.26-h9925aae_0 --> 2025.06.26-h9925aae_1
  s2n                                     1.5.21-h7ab7c64_0 --> 1.5.22-h96f233e_0
  scikit-learn                        1.7.0-py312h7a48858_1 --> 1.7.1-py312h4f0b9e3_0
  scipy                              1.15.2-py312ha707e6e_0 --> 1.16.0-py312hf734454_0
  sqlite                                  3.50.2-hb7a22d2_0 --> 3.50.3-heff268d_0
  statsmodels                        0.14.4-py312h8b63200_2 --> 0.14.5-py312h8b63200_0
  types-python-date~            2.9.0.20250516-pyhd8ed1ab_0 --> 2.9.0.20250708-pyhd8ed1ab_0
  typing-extensions                       4.14.0-h32cad80_0 --> 4.14.1-h4440ef1_0
  typing_extensions                     4.14.0-pyhe01879c_0 --> 4.14.1-pyhe01879c_0
  xarray                              2025.7.0-pyhd8ed1ab_0 --> 2025.7.1-pyhd8ed1ab_0
  xcdat                                  0.9.0-pyhd8ed1ab_0 --> 0.9.1-pyhd8ed1ab_0
  xskillscore                           0.0.26-pyhd8ed1ab_1 --> 0.0.27-pyhd8ed1ab_0

Proceed ([y]/n)? y

And I can still import xcdat (without triggering a numba error)!

I don't know why the update worked today, but would have failed 2 days ago.

• Maybe it is related to what was updated today in the base environment?
• xcdat went from 0.9.0 to 0.9.1 since my initial report, but 0.9.1 was already available 2 days ago
• the update worked correctly in a cleaned and a non-cleaned environment

The good thing is that I was able to update my existing environment (with lots of packages) without having to re-create it from scratch. The less good thing is that apparently all I had to do was be patient and wait, after getting the weird downgrade everything suggested by conda. Trouble is that you never know how long you need to wait

On the other hand, I got some side effects coming from my updating my complex environment to numy>=2 with no constraints and getting the patched numba only after using the proper constraint. Maybe re-creating a new environment would have worked directly

I will keep on updating cluster O and cluster S regularly, and also try to experiment with re-creating environments without the multiple users having to change anything to keep on working