I am using the below Python code to get data from a remote Redshift database. I am usi

You can try something like. <div class="highlight highlight-source-r notranslate p

Make SSH tunnel to query a remote redshift database in R about ssh HOT 1 OPEN

fissehab commented on July 17, 2024

Make SSH tunnel to query a remote redshift database in R

from ssh.

Comments (1)

FlorianSchwendinger commented on July 17, 2024

You can try something like.

start_tunnel_process <- function(user, server, port, stderr = nullfile()) {
    proc <- callr::r_bg(
        function(user, server, port) {
            ssh_host <- paste(user, server, sep = "@")
            while (TRUE) {
                ssh_session <- ssh::ssh_connect(ssh_host, keyfile = "~/.ssh/id_rsa", verbose = FALSE)
                ssh::ssh_tunnel(ssh_session, port = port, target = "localhost:27017")
                ssh::ssh_disconnect(ssh_session)
            }
        },
        args = list(user, server, port),
        stdout = nullfile(),
        stderr = stderr
    )
    proc
}

I currently evaluating this for my use case with mongolite.

url <- sprintf("mongodb://%s:%s@%s:%s", user, password, host, as.integer(port))
con <- mongo(collection, db = database, url = url)
con$find()
con$disconnect()

You need the while loop since after you disconnect the tunnel will close.
The while will create a new tunnel. This also means it will fail, if the time between two connections is to short for the 2nd process to create a new tunnel. You can hack this by putting a fault tolerant loop around your connect function.

for (i in seq_len(100)) {
    status <- try(con <- mongo(collection, db = database, url = url), silent = TRUE)
    if (!inherits(status, "try-error")) break
    Sys.sleep(0.2)
}