24

I'm using the following code in RSelenium to open a browser. After I close the browser, or even close the handler by running remDr$close(), the port is still in use. I have to go to the terminal and manually kill the process so that the same port becomes available. Is there any automated way such that RSelenium makes the port free after it finishes scraping?

So here is my code:

library(RSelenium)
rD <- rsDriver(verbose = FALSE,port=4444L)
remDr <- rD$client
remDr$close()

Thanks

Salim B
  • 1,623
  • 18
  • 25
Mohammad
  • 1,213
  • 1
  • 14
  • 30

5 Answers5

24

The process is composed of two parts a server (the Selenium Server) and a client (the browser you initiate). The close method of the remoteDriver class closes the client (the browser). The server also needs to be stopped when you are finished.

To stop the server when you are finished:

library(RSelenium)
rD <- rsDriver(verbose = FALSE,port=4444L)
remDr <- rD$client
remDr$close()

Now either explicitly stop the server:

rD$server$stop()

or if the rD object is removed the server will be stopped upon garbage collection:

library(RSelenium)
rD <- rsDriver(verbose = FALSE,port=4444L)
remDr <- rD$client
remDr$close()
rm(rD)
gc()
jdharrison
  • 28,335
  • 4
  • 67
  • 86
  • 5
    hello. tried both of these ac couple of times, no go. the broswer closes, but when i try to restart the rD object, i always get error that the port 4567) is already in use. only quitting the R session altogether releases it. – Guy Manova Feb 08 '20 at 23:03
  • 1
    I have the same issue as @GuyManova. – syre Mar 12 '20 at 09:51
  • So adding my 2 cents here. Seems like the solution doesn't work for the updated Rselenium package. Used the solution recommended by @Seb_ISU below , found on the thread here : https://github.com/ropensci/RSelenium/issues/228 – Bogdan Aug 13 '20 at 13:04
14

I did not have issues until recently. What worked for me is to use the solution above and as per the solution in this thread to add a line to kill the Java instance(s) inside RStudio.

remDr$close()
driver$server$stop()
rm(driver, remDr)
gc()

system("taskkill /im java.exe /f", intern=FALSE, ignore.stdout=FALSE)
Seb_ISU
  • 265
  • 4
  • 11
7

The command:

system("taskkill /im java.exe /f", intern=FALSE, ignore.stdout=FALSE)

will free all the ports.

If you want to free a particular port, you can do this:

#get the PID of the process you launched

pid <- driver$server$process$get_pid()

#pasting this PID in the following command (will kill all the child processes as well, closes the browser as well)

system(paste0("Taskkill /F /T" ," /PID ", pid))
Yugesh
  • 71
  • 1
  • 3
3

One way to avoid this problem is to use free_port() to find a free port (rather than specifying it manually)

library(netstat)
rsDriver(verbose = FALSE, port=free_port())
stevec
  • 15,490
  • 6
  • 67
  • 110
2

What worked for me is not calling stop at all and only calling close.

rD <- rsDriver(port = 4444L)
remDr <- rD[["client"]]
remDr$close()
rm(rD)
gc()

EDIT: Nevermind - this worked last week several times and then hasn't worked again.

boutitdoubtit
  • 31
  • 3