conda
Conda gets stuck / hangs at "Collecting package metadata (current_repodata.json)" using create or install or update
Checklist
- [x] I added a descriptive title
- [X] I searched open reports and couldn't find a duplicate
What happened?
Conda does not move past the "Collecting package metadata (current_repodata.json) : " stage when I run the command with options such as create, install or update. I have left it running for a few hours, so I presume it is stuck in an infinite loop somewhere(?).
I have been using conda for a while now, and this issue has popped up recently when I tried to create a new environment. I have uninstalled and freshly installed miniconda3 (PY3.9), yet the issue persists. I have cleaned up all conda files before reinstalling as is stated in the official documentation.
So, I am not sure what suddenly caused this issue. Thanks in advance for the help.
Conda Info
active environment : base
active env location : /home/kenja/miniconda3
shell level : 1
user config file : /home/kenja/.condarc
populated config files : /home/kenja/.condarc
conda version : 4.12.0
conda-build version : not installed
python version : 3.9.12.final.0
virtual packages : __cuda=11.6=0
__linux=5.15.0=0
__glibc=2.31=0
__unix=0=0
__archspec=1=x86_64
base environment : /home/kenja/miniconda3 (writable)
conda av data dir : /home/kenja/miniconda3/etc/conda
conda av metadata url : None
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 : /home/kenja/miniconda3/pkgs
/home/kenja/.conda/pkgs
envs directories : /home/kenja/miniconda3/envs
/home/kenja/.conda/envs
platform : linux-64
user-agent : conda/4.12.0 requests/2.27.1 CPython/3.9.12 Linux/5.15.0-46-generic ubuntu/20.04.4 glibc/2.31
UID:GID : 1000:1000
netrc file : None
offline mode : False
Conda Config
==> /home/kenja/.condarc <==
auto_activate_base: False
Conda list
# packages in environment at /home/kenja/miniconda3:
#
# Name Version Build Channel
_libgcc_mutex 0.1 main defaults
_openmp_mutex 4.5 1_gnu defaults
brotlipy 0.7.0 py39h27cfd23_1003 defaults
ca-certificates 2022.3.29 h06a4308_1 defaults
certifi 2021.10.8 py39h06a4308_2 defaults
cffi 1.15.0 py39hd667e15_1 defaults
charset-normalizer 2.0.4 pyhd3eb1b0_0 defaults
colorama 0.4.4 pyhd3eb1b0_0 defaults
conda 4.12.0 py39h06a4308_0 defaults
conda-content-trust 0.1.1 pyhd3eb1b0_0 defaults
conda-package-handling 1.8.1 py39h7f8727e_0 defaults
cryptography 36.0.0 py39h9ce1e76_0 defaults
idna 3.3 pyhd3eb1b0_0 defaults
ld_impl_linux-64 2.35.1 h7274673_9 defaults
libffi 3.3 he6710b0_2 defaults
libgcc-ng 9.3.0 h5101ec6_17 defaults
libgomp 9.3.0 h5101ec6_17 defaults
libstdcxx-ng 9.3.0 hd4cf53a_17 defaults
ncurses 6.3 h7f8727e_2 defaults
openssl 1.1.1n h7f8727e_0 defaults
pip 21.2.4 py39h06a4308_0 defaults
pycosat 0.6.3 py39h27cfd23_0 defaults
pycparser 2.21 pyhd3eb1b0_0 defaults
pyopenssl 22.0.0 pyhd3eb1b0_0 defaults
pysocks 1.7.1 py39h06a4308_0 defaults
python 3.9.12 h12debd9_0 defaults
readline 8.1.2 h7f8727e_1 defaults
requests 2.27.1 pyhd3eb1b0_0 defaults
ruamel_yaml 0.15.100 py39h27cfd23_0 defaults
setuptools 61.2.0 py39h06a4308_0 defaults
six 1.16.0 pyhd3eb1b0_1 defaults
sqlite 3.38.2 hc218d9a_0 defaults
tk 8.6.11 h1ccaba5_0 defaults
tqdm 4.63.0 pyhd3eb1b0_0 defaults
tzdata 2022a hda174b7_0 defaults
urllib3 1.26.8 pyhd3eb1b0_0 defaults
wheel 0.37.1 pyhd3eb1b0_0 defaults
xz 5.2.5 h7b6447c_0 defaults
yaml 0.2.5 h7b6447c_0 defaults
zlib 1.2.12 h7f8727e_1 defaults
Additional Context
In above three cases, the process is stuck at Collecting package metadata stage as mentioned previously.
In this last case, it is stuck after Solving environment stage.
Please note no new environment or library were successfully installed in any case as none of the above 4 processes ended.
But, when I tried to kill the process using Ctrl+Z, the output was the same in all 4 cases -
I can think of several things to try and see if we can get more info.
- Enable verbosity with either
-vvvor settingCONDA_VERBOSITY=3and let's see if we can infer where it's getting stuck (fetching, loading or post-processing) - Try skipping
current_repodata.jsonby adding the flag--repodata-fn=repodata.json -
conda clean --allto remove all index caches - Use the
--offlineflag to see if the error comes from network issues.
Let us know!
@dasupradyumna did you manage to get past this?
@jaimergp I tried all the steps you mentioned, skipping repodata, conda clean --all, the --offline flag... all result in the below.
This is my output after about 15 minutes, i.e. same as output at 3s:
root@e0a380ae66c9:/sd# conda env create -vvv -f environment.yaml
DEBUG conda.gateways.logging:set_verbosity(246): verbosity set to 3
DEBUG conda.core.solve:solve_final_state(258): solving prefix /opt/conda/envs/ldm
specs_to_remove: frozenset()
specs_to_add: frozenset({MatchSpec("cudatoolkit=11.3"), MatchSpec("pip=20.3"), MatchSpec("numpy=1.19.2"), MatchSpec("torchvision=0.12.0"), MatchSpec("python=3.8.5"), MatchSpec("pytorch=1.11.0"), MatchSpec("git")})
prune: False
Collecting package metadata (repodata.json): ...working...
Hoping for some help.
Sorry for disappearing, it was an important machine for running processes, but no stored data. So, I just uninstalled and reinstalled a fresh Ubuntu OS. Now, conda works as expected. I will leave this issue open since @webel seems to be still facing the same issue, and I presume a few others might be too. If not, I will close it.
Hi, I’m having the same bug here (exact same output, reproduced skipping repodata, conda clean --all, the --offline flag etc.), if you need more input on this.
A bit more info: I tried running it in pdb, and then Ctrl-C’ing to see where exactly it hangs, but it is totally unresponsive to Ctrl-C. had to kill it
EDIT: weirdly, trying to reinstall conda in another directory (through the Miniconda3-latest-Linux-x86_64.sh install script) hangs at the exact same step (Collecting package metadata (repodata.json))
EDIT 2: I’m on ubuntu like OP (18.04.5 LTS)
$ CONDA_VERBOSITY=3 conda create --name test
DEBUG conda.gateways.logging:set_verbosity(236): verbosity set to 3
DEBUG conda.core.solve:solve_final_state(258): solving prefix /home/tristan/miniconda3/envs/te
st
specs_to_remove: frozenset()
specs_to_add: frozenset()
prune: Null
Collecting package metadata (current_repodata.json): ...working... ^Z
[1]+ Stopped ( __add_sys_prefix_to_path; "$CONDA_EXE" $_CE_M $_CE_CONDA "$@"
)
$ kill %1
$ CONDA_VERBOSITY=3 conda create --name test --repodata-fn=repodat
a.json --offline
DEBUG conda.gateways.logging:set_verbosity(236): verbosity set to 3
DEBUG conda.core.solve:solve_final_state(258): solving prefix /home/tristan/miniconda3/envs/te
st
specs_to_remove: frozenset()
specs_to_add: frozenset()
prune: Null
Collecting package metadata (repodata.json): ...working... ^Z
[1]+ Stopped ( __add_sys_prefix_to_path; "$CONDA_EXE" $_CE_M $_CE_CONDA "$@"
)
Were the logs for that. But I’m ashamed to say that a system update (apt update / apt upgrade) and a reboot fixed it (should have tried it at first).
Still a mystery why this happened though, but i feel this is not a problem with conda, but with one of the things conda calls...
I had this error when I was working on a cluster, I forget to add CPU/GPU when i was creating the environment. See if you have enough ram when installing using conda.
Facing exactly this. It just gets stuck after "being done".
Preparing transaction: done
Verifying transaction: done
Executing transaction: done
#
# To activate this environment, use
#
# $ conda activate vidgen
#
# To deactivate an active environment, use
#
# $ conda deactivate
prompt is not returned, one core at 100%:
sadmin 1486 100 0.0 185544 52360 pts/0 R 13:11 9:35 /home/sadmin/miniconda3/bin/python -c from multiprocessing.spawn import spawn_main; spawn_main(tracker_fd=6, pipe_handle=8) --multiprocessing-fork
sadmin 3647 0.0 0.0 16260 10628 pts/0 S 13:18 0:00 /home/sadmin/miniconda3/bin/python -c from multiprocessing.resource_tracker import main;main(5)
sadmin 3648 0.1 0.0 185976 51488 pts/0 D 13:18 0:00 /home/sadmin/miniconda3/bin/python -c from multiprocessing.spawn import spawn_main; spawn_main(tracker_fd=6, pipe_handle=8) --multiprocessing-fork
sadmin 3744 2.2 0.1 739208 180544 pts/0 S+ 13:18 0:03 /home/sadmin/miniconda3/bin/python /home/sadmin/miniconda3/bin/conda create -n vidgen python=3.9
sadmin 3746 0.0 0.0 16260 10704 pts/0 S+ 13:18 0:00 /home/sadmin/miniconda3/bin/python -c from multiprocessing.resource_tracker import main;main(5)
sadmin 3747 0.1 0.0 185976 51352 pts/0 D+ 13:18 0:00 /home/sadmin/miniconda3/bin/python -c from multiprocessing.spawn import spawn_main; spawn_main(tracker_fd=6, pipe_handle=8) --multiprocessing-fork
16 cores of ryzen, 96GB ram, 1gb of ram used... Completely fresh, new installation of ubuntu server 22.04.
If killed, this is the trace:
^CError in atexit._run_exitfuncs:
Traceback (most recent call last):
File "/home/sadmin/miniconda3/lib/python3.9/multiprocessing/popen_fork.py", line 27, in poll
pid, sts = os.waitpid(self.pid, flag)
KeyboardInterrupt
It gets stuck REALLY HARD:
A new ubuntu install, using the very same installer and basically doing everything the same way - works perfectly.
The only difference here, is that the nvidia drivers are not installed yet.
This remains a big mystery.
After an hour of using it and installing torch, cuda stuff, eventually it stops working.
trace of the sudden death:
rmdir("/home/sadmin/miniconda3/envs/textgen/.condatmp") = 0
openat(AT_FDCWD, "/home/sadmin/miniconda3/envs/textgen/.condatmp", O_RDONLY|O_NONBLOCK|O_CLOEXEC|O_DIRECTORY) = -1 ENOENT (No such file or directory)
lstat("/home/sadmin/miniconda3/envs/textgen/.condatmp", 0x7ffd2fec82a0) = -1 ENOENT (No such file or directory)
stat("/home/sadmin/miniconda3/envs/textgen/.condatmp", 0x7ffd2fec82a0) = -1 ENOENT (No such file or directory)
write(1, "#\n# To activate this environment"..., 143#
# To activate this environment, use
#
# $ conda activate textgen
#
# To deactivate an active environment, use
#
# $ conda deactivate
) = 143
write(1, "\n", 1
) = 1
wait4(6907, 0x7ffd2fec9554, WNOHANG, NULL) = 0
kill(6907, SIGTERM) = 0
wait4(6907, 0x7ffd2fec9554, WNOHANG, NULL) = 0
getpid() = 6896
wait4(6907,
I have no idea how is it related, in my case, the issue was a device (RTX 3090) not connecting properly via PCI passthrough. Absolutely no idea how is it related, but once I fixed that (or removed the device) the hang and 100% CPU core usage in kernel space stopped.
Hey @ThatCoffeeGuy, this is probably related to the CUDA detection plugin. If you can reproduce again, try if CONDA_OVERRIDE_CUDA=10 changes anything.
I think I got a very similar situation as you described, the only difference is I got AttributeError: 'Namespace' object has no attribute 'func' when running conda -vvv.
I didn't solve the problem yet
