This is the constructor used to create custom cluster functions. Note that some standard implementations for TORQUE, Slurm, LSF, SGE, etc. ship with the package.

makeClusterFunctions(name, submitJob, killJob = NULL, listJobsQueued = NULL,
  listJobsRunning = NULL, array.var = NA_character_,
  store.job.collection = FALSE, store.job.files = FALSE,
  scheduler.latency = 0, fs.latency = NA_real_, hooks = list())

Arguments

name

[character(1)]
Name of cluster functions.

submitJob

[function(reg, jc, ...)]
Function to submit new jobs. Must return a SubmitJobResult object. The arguments are reg (Registry) and jobs (JobCollection).

killJob

[function(reg, batch.id)]
Function to kill a job on the batch system. Make sure that you definitely kill the job! Return value is currently ignored. Must have the arguments reg (Registry) and batch.id (character(1) as returned by submitJob). Note that there is a helper function cfKillJob to repeatedly try to kill jobs. Set killJob to NULL if killing jobs cannot be supported.

listJobsQueued

[function(reg)]
List all queued jobs on the batch system for the current user. Must return an character vector of batch ids, same format as they are returned by submitJob. Set listJobsQueued to NULL if listing of queued jobs is not supported.

listJobsRunning

[function(reg)]
List all running jobs on the batch system for the current user. Must return an character vector of batch ids, same format as they are returned by submitJob. It does not matter if you return a few job ids too many (e.g. all for the current user instead of all for the current registry), but you have to include all relevant ones. Must have the argument are reg (Registry). Set listJobsRunning to NULL if listing of running jobs is not supported.

array.var

[character(1)]
Name of the environment variable set by the scheduler to identify IDs of job arrays. Default is NA for no array support.

store.job.collection

[logical(1)]
Flag to indicate that the cluster function implementation of submitJob can not directly handle JobCollection objects. If set to FALSE, the JobCollection is serialized to the file system before submitting the job.

store.job.files

[logical(1)]
Flag to indicate that job files need to be stored in the file directory. If set to FALSE (default), the job file is created in a temporary directory, otherwise (or if the debug mode is enabled) in the subdirectory jobs of the file.dir.

scheduler.latency

[numeric(1)]
Time to sleep after important interactions with the scheduler to ensure a sane state. Currently only triggered after calling submitJobs.

fs.latency

[numeric(1)]
Expected maximum latency of the file system, in seconds. Set to a positive number for network file systems like NFS which enables more robust (but also more expensive) mechanisms to access files and directories. Usually safe to set to NA to disable the heuristic, e.g. if you are working on a local file system.

hooks

[list]
Named list of functions which will we called on certain events like “pre.submit” or “post.sync”. See Hooks.

See also