0

We have server behind proxy and we want this server to be able to run commands such as:

  • python: pip install module
  • R: install.packages("fortunes")
  • ...

Simply to install packages from these sources. Since we are behind proxy, we cannot install these unless the proxy has them whitelisted (otherwise the proxy probihits the connection between the server and wherever the package resides).

My question is: what should we whitelist to be able to run these commands?

I am not sure how the package websites actually works (whether they store the packages themselves or it is just the index and the actual packages resides on other domains/hostnames/...). I believe pypi is quite friendly here (packages are actually found there), but CRAN or Maven = don't know. We are running Spark servers, so our primary concerns are python, R, Java or Scala libraries/packages.

mLC
  • 663
  • 10
  • 22

1 Answers1

0

Maven: is actually storing packages. Regarding mirroring, see this answer. It also contains the url of the central repository.

Pypi: From the documentation on how to upload a package to the index, it seems like it is also physically storing the packages.

CRAN: also hosts the packages. There are several mirrors, you will need to whitelist one you want to use

You might want to consider setting up an internal mirror where you put your dependencies once, and then don't need to go to the outside internet.

kutschkem
  • 7,826
  • 3
  • 21
  • 56