I'd like to install the pymongo library but I'm getting the following error:
(C:\Users\xxxxxxx\AppData\Local\Continuum\anaconda3) C:\Users\xxxxxxx>
conda install -c anaconda pymongo
Fetching package metadata ...
CondaHTTPError: HTTP 000 CONNECTION FAILED for url <https://conda.anaconda.org/a
naconda/win-64/repodata.json>
Elapsed: -
An HTTP error occurred when trying to retrieve this URL.
HTTP errors are often intermittent, and a simple retry will get you on your way.
ConnectTimeout(MaxRetryError("HTTPSConnectionPool(host='conda.anaconda.org', por
t=443): Max retries exceeded with url: /anaconda/win-64/repodata.json (Caused by
ConnectTimeoutError(<urllib3.connection.VerifiedHTTPSConnection object at 0x000
00000054D6128>, 'Connection to conda.anaconda.org timed out. (connect timeout=9.
15)'))",),)
Steps taken to resolve:
1. Update C:\Users\\xxxxxxx\.condarc file with the following:
channels:
- defaults
ssl_verify: false
proxy_servers:
http: http://sproxy.fg.xxx.com:1000
https: https://sproxy.fg.xxx.com:1000
2. (C:\Users\xxxxxxx\AppData\Local\Continuum\anaconda3) C:\Users\xxxxxxx>
conda config --set ssl_verify False
Additional Info:
(C:\Users\xxxxxxx\AppData\Local\Continuum\anaconda3) C:\Users\xxxxxxx>
conda info
Current conda install:
platform : win-64
conda version : 4.3.27
conda is private : False
conda-env version : 4.3.27
conda-build version : 3.0.22
python version : 3.6.2.final.0
requests version : 2.18.4
config file : C:\Users\xxxxxxx\.condarc
netrc file : None
offline mode : False
user-agent : conda/4.3.27 requests/2.18.4 CPython/3.6.2 Windows/7 W
indows/6.1.7601
administrator : False
A number of posts online simply reinstalled Anaconda, any other options apart from a fresh install?
This works a charm:-
Just copy these:-
libcrypto-1_1-x64.dll
libssl-1_1-x64.dll
from D:\Anaconda3\Library\bin to D:\Anaconda3\DLLs.
Execute the following command in the cmd prompt/terminal:
conda config --set ssl_verify no
I try to create a virtual env with python 2.7 with anaconda, the base env is python 3.7. I encounter the exactly same problem. It turns out that there isn't such problem with other virtual envs with python 3.7 or 3.6.
This post works perfectly to solve my problem on win7 with anaconda prompt.
It basically says you need to add the following directories into your user environment path in windows (go to Start and type in: View Advanced System Settings, then select Environmental Variables: then select Path and click Edit: finally you can click New and add a path):
C:\your_directory_to_anaconda3\Anaconda3\Scripts
C:\your_directory_to_anaconda3\Anaconda3\
C:\your_directory_to_anaconda3\Anaconda3\Library\bin -- This is the directory for openssl
If you added conda to your PATH variables, remove it and use the "Anaconda Prompt". This solved the problem for me.
See: https://github.com/conda/conda/issues/8046#issuecomment-450582208
I faced this issue when I tried to create environment. I solved it by first activating conda base environment by using:
conda activate base
then I created the environment
conda create -n myenv python=3.7
Check the proxy URL
Verify .condarc file
For me, the problem was with the indentation in the .condarc file.
proxy_servers:
http: http://testproxy:8080
https: https://testproxy:8080
My authenticated proxy server is configured with a domain whitelist for massive and repeated downloads so root or local sudoer doesn't need to be authenticated.
Adding conda.anaconda.org is not enough as this repo redirect its traffic to amazonaws.com.
In my case, adding ".amazonaws.com" to the whitelist solved the issue.
The issue was resolved by adding a username and password to file C:\Users\xxxxx.condarc
channels:
- defaults
ssl_verify: false
proxy_servers:
http: http://xxxxx:password#sproxy.fg.abc.com:yyyy
https: https://xxxxx:password#sproxy.fg.abc.com:yyyy
I had the same problem on Windows 10-64 bit and intuitively installed the 64-bit version of miniconda. However, it results in exactly the same error above. Installing 32 bit conda installer has resolved the issue
Before installing some package (pydicom) the installation run just fine. After
it I tried to install matplotlib, but I got the same error as yours.
I tried conda config --set ssl_verify no but it didn't solve the problem so I set it again to true.
Fortunately, I had a virtual environment where I installed my packages. I closed all Anaconda prompts and tried in a new test environment. Magically, the install worked. I came back to my original virtual env and run the install again, and it worked!.
It might be that I just had to wait for some time before I could use conda install again.
One other thing I could do is remove the package that caused the problem, but I didn't have the chance to try it. If it has anything to do with some virtual environments not being affected, then one possible way to guard against this is to clone the environment before installing any new package.
Edit: I tried the same solution but It did not work. But instead of showing the error immediately, it asks me whether I want to proceed. I deactivated the env, and re-opened anaconda prompt, then did the same steps as above and worked again.
I also had the Same Issue, I resolved by installing 32 bit Anaconda Installer.
Which resolved the CondaHTTPError: HTTP 000 CONNECTION, on Windows 64 bit.
I faced this issues after "conda clean -a" on win-64.
Activating and deactivating existing conda env resolved the issue.
You might need to upgrade your openssl installation
You can download it here (Try the latest version):
https://slproweb.com/products/Win32OpenSSL.html
Source:
https://github.com/ContinuumIO/anaconda-issues/issues/6424#issuecomment-464660808
My issue was simply not running the conda init command prior to attempting to create an environment.
Came across the CondaHTTP Connection error after installing Anaconda environment on a new Windows 10 computer. I tried virtually all the recommendations above unsuccessfully! Looking up the Anaconda archives ( https://repo.continuum.io/archive/ ), I downloaded the immediate previous release .... and on installation and rebooting my PC, all is now wellscreenshot of release
In short - installing Microsoft Visual C++ Redistributable for Visual Studio solved my problem.
In more detail: upon trying a suggested solution of installing a new version of OpenSSL, the installation process told me I was missing a dependency - the Visual Studio Redistributable package. The installer led me to a direct download page of the 2017 version. I can't find that page now, but the official release of 2019 can be found here, and should work as well (found under Other Tools and Frameworks).
Uninstalling and reinstalling anaconda for all users (instead, of current user only, requires admin privileges) and activating the option to add Anaconda to PATH during the installation process, fixed the issues for me.
Thank you everyone for your responses. In my case, I found out that my Kaspersky Internet Security was blocking it the whole time. The moment I quit the application all applications were downloaded. Please check your firewall settings before trying all the above options.
I tried all of these solutions and none worked for me. After running the command
conda config --remove-key channels
in the Anaconda Prompt, everything started working for me on my next attempt.
Adding that I had the same problem on ubuntu on WSL. None of the solutions worked for me, until I realized I was working on WSL version 1 (I thought I'd already upgraded). Upgrading from WSL 1 to WSL 2 solved the problem for me.
Running following these two commands worked for me.
conda config --remove-key proxy_servers
conda clean --source-cache
I'd tried all of the advice on this and many other webpages.
In the end I broadcast a "help me Obiwan Kenobi, you're my only hope message" to a large group of people at work and one of them who used python all the time was able to help me
The trick was to set several windows environmental variables
CURL_CA_BUNDLE
REQUESTS_CA_BUNDLE
SSL_CERT_FILE
To my company's root certificate (a .cer or .crt) which I had downloaded to a spot on my disk
You may also need to add (in my case)
C:\Users\kdalbey\Anaconda3\Scripts
(or your particular \Anaconda3\Scripts) to your path.
And then I set proxies just for good measure
note I previously copied libcrypto-1_1-x64.dll, libcrypto-1_1-x64.pdb, libssl-1_1-x64.dll, libssl-1_1-x64.pdb from anaconda3\Library\bin to anaconda3\DLLs so that could be part of the secret sauce
and it didn't work until I killed and restarted anaconda-navigator
Two steps to deal with this error.
The Anaconda prompt configures the path, to include all the necessary executable files (for instance Library\bin - On Windows, launch it with admin permission). So that you need to use it to execute conda :
Update conda with conda update conda
Exit my proxy software which solved the issue.
I would like summarize some of the proposed answers in this post and propose my experience on that. As it can be understood from the error explanation, the error is related to the connection and I strongly believe that no need to uninstall and reinstall anything if the real cause of the problem be known. My problem gone away after the system powered off and powered on again one day later. So, some possible causes and their solutions (these solutions could be tested in order based on the written bulleted order) could be as follows:
Crash in anaconda prompt:
Probable solutions:
Deactivating and activating the environment, without removing all packages or …, or
Closing/reopening the prompt (Michael Heidelberg) or
Using cmd.exe instead, perhaps
Non-responsiveness of the anaconda site:
Massive site traffics related probable issues, that could be the reason of non-responsiveness or to temporary block some IPs
Probable solutions:
Retying as recommended in the error: HTTP errors are often intermittent, and a simple retry will get you on your way. It solves my problem sometimes. or
Activate or deactivate VPNs or Proxies (like use in .condarc; see: Github sroder, Nandhan Thiravia, Vinod Sangale, Peter Lucas, Sunding Wei).
Try after a while if you have time
System firewall block the site:
That might be happened by activating and deactivating of VPNs, repeatedly or by some other works
Probable solutions:
Finding the issue in system firewall and allowing the connection in the firewall settings (ScienceJedi, Github)
Reboot, perhaps
If the aforementioned ways didn't solve the problem, testing the related answers in the following order:
Add ...\Anaconda3\Scripts, ...\Anaconda3\, and ...\Anaconda3\Library\bin to the path (talentcat, skerjj, Victor Ochieng, jankap), perhaps need a reboot after (lightarrow)
Copying libcrypto-1_1-x64.dll and libssl-1_1-x64.dll from D:\Anaconda3\Library\bin into D:\Anaconda3\DLLs (Swapnil)
I think it could be used in the first step because It is unlikely to be cause of any other problem. The reason I didn't mention this at the beginning is that the developers could placed these files in that directory during installation, too, in default, but they didn't; perhaps it had some reasons (Github).
Note: these files are for Python >3, and I didn't find them for Python 2. Perhaps they have another names.
It must be said that my problem didn't solve by this solution.
Keep your SSL stack up-to-date (kamal dua, Anaconda troubleshooting, update openssl, Abdulrahman Bres, Update to openssl 1.1.1)
I didn't recommend it at first because Its not a good idea to unset ssl verification unless you know what you are doing (Pratyush comment) and somewhere I read that it couldn't return to True again.
It must be said that my problem didn't solve by this solution, too.
conda config --set ssl_verify false
conda update openssl ca-certificates certifi
I am trying to follow this tutorial, in order to install the Nix package manager in my home directory instead of /nix.
I am doing the PRoot installation (see 2. in tutorial). At the end, the
tutorial proposes to be smart in Building native packages section, to be
able to run packages without PRoot:
To run packages natively (without PRoot) they have to be build from source because all paths to the nix store are hard-coded. It is simple, really:
mkdir $HOME/nix
nix-channel --update
env NIX_STORE_DIR=$HOME/nix nix-env -i nix
And now your Nix store gets built up using the new paths. The built binaries can be run directly from there.
I did that, but I don't see how it frees me from PRoot. If I don't do the /nix mounting point with PRoot, nothing works (no nix-env executable,
I can't install new packages).
Should this NIX_STORE_DIR environment variable be put in my .bashrc ?
It seems I always need to run PRoot because ~/.nix-profile points to
a /nix/... directory:
.nix-profile -> /nix/var/nix/profiles/default
There are more steps in the tutorial (5., 6.) - should I follow them ? It seems they apply only in case of using the manual installation (step 4.),
although it is not explicit.
Any help would be appreciated :)
For anyone stumbling on this old question: there is no currently supported way to install Nix without root. The above wiki was moved to https://nixos.wiki/wiki/Nix_Installation_Guide . It may well be out of date. PRoot could work, but even then, rebuilding the whole store at a different path is not a good idea, not the least because the binary caches won't help and you'll need to build everything.
I suggest trying Nix in a virtual machine or cloud server.
Future people from Google, it's still unsupported but does work. Script here that installs a couple dependencies, builds a temporary Nix, and uses that to install a proper version in your directory of choice.
In the R FAQ section 4.6 (Package TclTk does not work) I found the following sentence:
... although they [missing Tcl/tk packages] may be downloaded via the Teacup facility
What is "teacup"? How can I install and use it?
I am using RStudio running on Ubuntu Linux and Windows 7.
Teacup is a program that ships as part of ActiveTcl, a commercial zero-cost distribution of Tcl (and Tk and many other packages) for various platforms. It does package management, looking after the key part that is download, installation and upgrading of packages from a remote repository. It is not open source, though Tcl itself is (as are the majority of packages that aren't single-company-specific).
If you've got it installed, you use these commands from a shell:
teacup update-self
teacup update
Depending on where your Tcl installation is, you might need to elevate privileges to make these command calls work. How you do this is platform-dependent; on Unix it's usually simplest to use sudo for each of the commands, whereas on Windows it is probably easier to create an elevated command shell and run inside that.
Depending on your site, you might need to configure a web proxy with teacup proxy. Try without first.
If you're using a non-ActiveTcl installation but you have an ActiveTcl installation present, you can still use teacup. You just need to use teacup link to connect that Tcl installation to the teacup local repository. This is slightly more complex because you can have multiple repositories on the one system (though I've never needed that).
First, you find where the repository is:
teacup default
Then you need to link the shell to the repository:
teacup link make $PATH_FROM_TEACUP_DEFAULT $LOCATION_OF_TCLSH_TO_LINK
Making this work with R Studio will be a matter of determining which Tcl installation it is using. If it's already an ActiveTcl, you just need the first part of this answer. Otherwise, you need the second part as well. Also note that pretty much requires that you be using either Tcl 8.5 or 8.6; there are no guarantees for older, unsupported versions.
For my system: Ubuntu 12.04 and R 3.03, whenever I install a custom package in R via
>install.packages()
the package is installed by default to
/home/USER/R/x86_64-pc-linus-gnu-library/3.0/
as opposed to system-wide in
/usr/local/lib/R/site-library/
which is needed for shiny-server to work with that package.
My temporary solution is to copy the packages to the correct folder after the fact.
Question: How can I set the default install path from the start to avoid this problem?
Yes -- I consider this to be a misfeature and disable my per-user directory.
Moreover, I mostly use a script install.r (of which a version is an example in the littler package you can install as part of Ubuntu) which simple explicitly set the /usr/local/lib/R/site-library directory as the default. With a patch we got into R 3.0.2 or 3.0.3, normal user can write into the directory and will now create group-writeable directories so other users can update and overwrite -- just make everybody a member of the same group, say staff or admin. And then you don't even need sudo or root.
I have essentially answered this same question a few times here over the years (minus the shiny angle, which is not really relevant) so feel free to search for the other for more details, examples, ...
I would propose a different approach.
The problem is that shiny-server cannot find the packages that you install because it runs them as a different user which is called shiny. This user is created upon installation of shiny-server
The easiest (and safest IMHO) way to solve this is to just install the packages as the shiny user, using the following steps.
Set a password for the user using sudo passwd shiny, now enter and confirm a password of your choosing.
Switch to the shiny account using: su - shiny
Call up R using $ R (without sudo)
Install the required packages, in this case: `install.packages("shinydashboard")
Note that if you have rstudio-server installed on the same machine then you can perform steps 2-4 using that interface. Simply go the same domain/ip and use :8787 for the rstudio-server interface instead of :3838 for shiny-server.
Adapted from my answer here