Use R on HDFS With High Availability

You can use R on HDFS with the high availability option. The high availability option allows you to have two NameNodes for one HDFS in case of failure.

Install the following package:
```
install.packages("httr")
library(httr)
```

Run the following lines of code:

# To list all the URLs you want to try.
uris <- c('http://nn1.p01.client.saagie.io:50070/webhdfs/v1/', 'http://nn1.p02.client.saagie.io:50070/webhdfs/v1/',
'http://nn2.p01.client.saagie.io:50070/webhdfs/v1/', 'http://nn2.p02.client.saagie.io:50070/webhdfs/v1/')

# This function tries all URLs and filters out those that return a timeout or a bad status.
getNN <- function(uris) {
  op <- '?op=LISTSTATUS' # Operation to try. In this case list, the contents of a folder.
  res <- lapply(uris, function(uri) {
    # If timeout, it returns the status code 400, else it returns the status code of the request.
    status <- tryCatch( status_code(httr::GET(paste0(uri, op), timeout(1))), error=function(x) 400 )
    if(status != 200) return(NULL) # If the status code is invalid, returns NULL.
    uri # If the status is valid, returns the working URLs.
  })
  unlist(uris[lapply(res, length) > 0]) # To return all working URLs and filter the rest.
}

first_valid <- getNN(uris)[1] # Select the first working URL.