launch: Launches a persistent pbdR server on a cluster

In RBigData/launchr: Launch a distributed R server on a cluster from a remote R session

Launches a persistent pbdR server on a cluster

Description

Launches a persistent pbdR server on a cluster

Usage

launch(server, user = Sys.getenv("USER"), FUN = function(script, ...)
  script, port = 55555, verbose = FALSE, fn = list(lnode_file =
  ".pbdR_server.sh", pbs_file = ".pbdR_server.pbs", head_node_file =
  ".pbdR_server_hnode", server_log = ".pbdR_server.o"), ...)

Arguments

server	A string giving the hostname to run the pbdR server, resolvable on your local platform.
user	Username for server login.
FUN	A function that modifies the default ssh script. It is called with FUN(script, ...) so it has access to unmatched arguments in launch(). The default is the identity function.
port	The port to be used for communication with the server. Defaults to 55555.
verbose	If TRUE, print generated scripts before submission. A value > 1 will also prevent script submission.
fn	A list of file names that will be created on the server. Normally the defaults are fine.
...	Arguments for script details. Typical parameters include nodes The number of nodes to allocate for the server. npernode The number of MPI ranks (R sessions) to start per node. walltime A string "hh:mm:ss" giving the maximum length of time in hours:minutes:seconds before server timeout shutdown. account Account used for your cluster allocation. modules A vector of strings giving the modules to load before starting the server. rwd A string giving a working directory on server.

Value

A shell script as a vector of strings.

RBigData/launchr documentation built on April 28, 2022, 10:33 a.m.