Skip to contents

Project Status: Active – The project has reached a stable, usable state and is being actively developed. CRAN_Status_Badge CRAN RStudio mirror downloads

Secure Shell (SSH) Client for R

Installation

This package is available on CRAN and can be installed via:

Alternatively it can be installed from source using devtools:

remotes::install_github('ropensci/ssh')

Installation from source on MacOS or Linux requires libssh (the original libssh, not the unrelated libssh2 library). On Debian or Ubuntu use libssh-dev:

sudo apt-get install -y libssh-dev

On Fedora we need libssh-devel:

sudo yum install libssh-devel

On CentOS / RHEL we install libssh-devel via EPEL:

sudo yum install epel-release
sudo yum install libssh-devel

On OS-X use libssh from Homebrew:

brew install libssh

Using conda (need a conda R environment conda create -n Renv r-base r-essentials)

conda install --channel conda-forge r-ssh

If you have issues with the conda installation please submit an issue in conda-forge/r-ssh-feedstock

Getting Started

First create an ssh session by connecting to an SSH server. You can either use private key or passphrase authentication:

session <- ssh_connect("jeroen@dev.opencpu.org")

You can use the session in subsequent ssh functions below.

Run a command

Run a command or script on the host while streaming stdout and stderr directly to the client.

ssh_exec_wait(session, command = c(
  'curl -fOL https://cloud.r-project.org/src/contrib/Archive/jsonlite/jsonlite_1.5.tar.gz',
  'R CMD check jsonlite_1.5.tar.gz',
  'rm -f jsonlite_1.5.tar.gz'
))

If you want to capture the stdout/stderr:

out <- ssh_exec_internal(session, "R -e 'rnorm(100)'")
cat(rawToChar(out$stdout))

Using ‘sudo’

Note that the exec functions are non interactive so they cannot prompt for a sudo password. A trick is to use -S which reads the password from stdin:

out <- ssh_exec_wait(session, 'echo "mypassword!" | sudo -s -S apt-get update -y')

Be very careful with hardcoding passwords!

Uploading and Downloading via SCP

Upload and download files via SCP. Directories are automatically traversed as in scp -r.

# Upload a file to the server
file_path <- R.home("COPYING")
scp_upload(session, file_path)
# Download the file back and verify it is the same
scp_download(session, "COPYING", to = tempdir())
tools::md5sum(file_path)
tools::md5sum(file.path(tempdir(), "COPYING"))

Create a Tunnel

Opens a port on your machine and tunnel all traffic to a custom target host via the SSH server.

ssh_tunnel(session, port = 5555,target = "ds043942.mongolab.com:43942")

This function blocks while the tunnel is active. Use the tunnel by connecting to localhost:5555 from a separate process. The tunnel can only be used once and will automatically be closed when the client disconnects.

Disconnect

When you are done with the session you should disconnect:

If you forgot to disconnect, the garbage collector will do so for you (with a warning).